Advertisement
Mayo Clinic Proceedings Home

Consumption of Ultra-Processed Foods and Mortality: A National Prospective Cohort in Spain

  • Ruth Blanco-Rojo
    Affiliations
    Instituto Madrileño De Estudios Avanzados-alimentacion-Food Institute, Campus de Excelencia Internacional Universidad Autónoma de Madrid + Centro Superior de Investigaciones Científicas, Madrid, Spain
    Search for articles by this author
  • Helena Sandoval-Insausti
    Affiliations
    Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Centro de Investigación Biomedica en Red of Epidemiology and Public Health, Madrid, Spain

    Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA
    Search for articles by this author
  • Esther López-Garcia
    Affiliations
    Instituto Madrileño De Estudios Avanzados-alimentacion-Food Institute, Campus de Excelencia Internacional Universidad Autónoma de Madrid + Centro Superior de Investigaciones Científicas, Madrid, Spain

    Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Centro de Investigación Biomedica en Red of Epidemiology and Public Health, Madrid, Spain
    Search for articles by this author
  • Auxiliadora Graciani
    Affiliations
    Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Centro de Investigación Biomedica en Red of Epidemiology and Public Health, Madrid, Spain
    Search for articles by this author
  • Jose M. Ordovás
    Affiliations
    Instituto Madrileño De Estudios Avanzados-alimentacion-Food Institute, Campus de Excelencia Internacional Universidad Autónoma de Madrid + Centro Superior de Investigaciones Científicas, Madrid, Spain

    Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain

    U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, MA
    Search for articles by this author
  • Jose R. Banegas
    Affiliations
    Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Centro de Investigación Biomedica en Red of Epidemiology and Public Health, Madrid, Spain
    Search for articles by this author
  • Fernando Rodríguez-Artalejo
    Affiliations
    Instituto Madrileño De Estudios Avanzados-alimentacion-Food Institute, Campus de Excelencia Internacional Universidad Autónoma de Madrid + Centro Superior de Investigaciones Científicas, Madrid, Spain

    Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Centro de Investigación Biomedica en Red of Epidemiology and Public Health, Madrid, Spain
    Search for articles by this author
  • Pilar Guallar-Castillón
    Correspondence
    Correspondence: Address to Pilar Guallar-Castillón, MD, PhD, Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid, Arzobispo Morcillo, n 4, 28029 Madrid, Spain.
    Affiliations
    Instituto Madrileño De Estudios Avanzados-alimentacion-Food Institute, Campus de Excelencia Internacional Universidad Autónoma de Madrid + Centro Superior de Investigaciones Científicas, Madrid, Spain

    Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Centro de Investigación Biomedica en Red of Epidemiology and Public Health, Madrid, Spain

    Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins University, Baltimore, MD
    Search for articles by this author
Open AccessPublished:October 14, 2019DOI:https://doi.org/10.1016/j.mayocp.2019.03.035

      Abstract

      Objective

      To assess the prospective association between ultra-processed food consumption and all-cause mortality and to examine the effect of theoretical iso-caloric non-processed foods substitution.

      Patients and Methods

      A population-based cohort of 11,898 individuals (mean age 46.9 years, and 50.5% women) were selected from the ENRICA study, a representative sample of the noninstitutionalized Spanish population. Dietary information was collected by a validated computer-based dietary history and categorized according to their degree of processing using NOVA classification. Total mortality was obtained from the National Death Index. Follow-up lasted from baseline (2008-2010) to mortality date or December 31th, 2016, whichever was first. The association between quartiles of consumption of ultra-processed food and mortality was analyzed by Cox models adjusted for the main confounders. Restricted cubic-splines were used to assess dose-response relationships when using iso-caloric substitutions.

      Results

      Average consumption of ultra-processed food was 385 g/d (24.4% of the total energy intake). After a mean follow-up of 7.7 years (93,599 person-years), 440 deaths occurred. The hazard ratio (and 95% CI) for mortality in the highest versus the lowest quartile of ultra-processed food consumption was 1.44 (95% CI, 1.01-2.07; P trend=.03) in percent of energy and 1.46 (95% CI, 1.04-2.05; P trend=.03) in grams per day per kilogram. Isocaloric substitution of ultra-processed food with unprocessed or minimally processed foods was associated with a significant nonlinear decrease in mortality.

      Conclusion

      A higher consumption of ultra-processed food was associated with higher mortality in the general population. Furthermore, the theoretical iso-caloric substitution ultra-processed food by unprocessed or minimally processed foods would suppose a reduction of the mortality risk. If confirmed, these findings support the necessity of the development of new nutritional policies and guides at the national and international level.

      Trial Registration

      Abbreviations and Acronyms:

      ENRICA (Study on Nutrition and Cardiovascular Risk in Spain), HR (hazard ratio)
      Processing of food arose as a need to improve food availability, safety, digestibility, transportability, and storage life.
      • Ludwig D.S.
      Technology, diet, and the burden of chronic disease.
      • Wrangham R.
      The evolution of human nutrition.
      In the past decades, processing of food and food supplies have increased broadly providing ready-to-consume processed products that can be distributed all around the world.
      • Monteiro C.A.
      • Moubarac J.C.
      • Cannon G.
      • Ng S.W.
      • Popkin B.
      Ultra-processed products are becoming dominant in the global food system.
      Several food classifications have been proposed to quantify the nature and the extent of processed food intake when measuring it in populations.
      Food and Agriculture Organization of the United Nations
      Guidelines on the collection of information on food processing through food consumption surveys.
      The most extreme category corresponds to the ultra-processed foods, which are formulations made mostly or entirely from substances derived from foods and additives, with little if any whole food.
      Food and Agriculture Organization of the United Nations
      Guidelines on the collection of information on food processing through food consumption surveys.
      Compared with the rest of the diet, ultra-processed foods often have a higher content of total fat, saturated fat, added sugars, and salt, along with a low amount of fiber and low vitamin density.
      • Luiten C.M.
      • Steenhuis I.H.
      • Eyles H.
      • Ni Mhurchu C.
      • Waterlander W.E.
      Ultra-processed foods have the worst nutrient profile, yet they are the most available packaged products in a sample of New Zealand supermarkets.
      • Monteiro C.A.
      • Cannon G.
      • Moubarac J.C.
      • Levy R.B.
      • Louzada M.L.C.
      • Jaime P.C.
      The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing.
      • Rauber F.
      • da Costa Louzada M.L.
      • Steele E.M.
      • Millett C.
      • Monteiro C.A.
      • Levy R.B.
      Ultra-processed food consumption and chronic non-communicable diseases-related dietary nutrient profile in the UK (2008-2014).
      In addition, they can contain neo-formed contaminants derived from physical, chemical, and biologic processes, along with substances from packaging and additives.
      • Martínez Steele E.
      • Monteiro C.A.
      Association between dietary share of ultra-processed foods and urinary concentrations of phytoestrogens in the US.
      • Scrinis G.
      • Monteiro C.A.
      Ultra-processed foods and the limits of product reformulation.
      Moreover, epidemiologic evidence has demonstrated that ultra-processed food consumption is associated with poorer diet quality.
      • Martínez Steele E.
      • Popkin B.M.
      • Swinburn B.
      • Monteiro C.A.
      The share of ultra-processed foods and the overall nutritional quality of diets in the US: evidence from a nationally representative cross-sectional study.
      • Moubarac J.C.
      • Batal M.
      • Louzada M.L.
      • Martinez Steele E.
      • Monteiro C.A.
      Consumption of ultra-processed foods predicts diet quality in Canada.
      On the other hand, ultra-processed foods are highly palatable, habit-forming, convenient, microbiologically safe, affordable, and aggressively advertised and marketed.
      • Monteiro C.A.
      Nutrition and health. The issue is not food, nor nutrients, so much as processing.
      • Moodie R.
      • Stuckler D.
      • Monteiro C.
      • et al.
      Lancet NCD Action Group
      Profits and pandemics: prevention of harmful effects of tobacco, alcohol, and ultra-processed food and drink industries.
      National household food purchase surveys and national dietary surveys have informed on ultra-processed food intake in some Western countries, such as the United States,
      • Baraldi L.G.
      • Martinez Steele E.
      • Canella D.S.
      • Monteiro C.A.
      Consumption of ultra-processed foods and associated sociodemographic factors in the USA between 2007 and 2012: Evidence from a nationally representative cross-sectional study.
      Canada,
      • Moubarac J.C.
      • Batal M.
      • Martins A.P.
      • et al.
      Processed and ultra-processed food products: consumption trends in Canada from 1938 to 2011.
      New Zealand,
      • Luiten C.M.
      • Steenhuis I.H.
      • Eyles H.
      • Ni Mhurchu C.
      • Waterlander W.E.
      Ultra-processed foods have the worst nutrient profile, yet they are the most available packaged products in a sample of New Zealand supermarkets.
      several European countries,
      • Julia C.
      • Martinez L.
      • Allès B.
      • et al.
      Contribution of ultra-processed foods in the diet of adults from the French NutriNet-Santé study.
      • Slimani N.
      • Deharveng G.
      • Southgate D.A.
      • et al.
      Contribution of highly industrially processed foods to the nutrient intakes and patterns of middle-aged populations in the European Prospective Investigation into Cancer and Nutrition study.
      and some South American countries, such as Brazil,
      • Costa Louzada M.L.
      • Martins A.P.
      • Canella D.S.
      • et al.
      Ultra-processed foods and the nutritional dietary profile in Brazil.
      and Chile.
      • Cediel G.
      • Reyes M.
      • da Costa Louzada M.L.
      • et al.
      Ultra-processed foods and added sugars in the Chilean diet (2010).
      It has been estimated that ultra-processed food intake is increasing, and it currently contributes 25%-50% of total energy intake.
      • Kelly B.
      • Jacoby E.
      Public Health Nutrition special issue on ultra-processed foods.
      These shifts in dietary habits have been parallel to an increase in chronic diseases, beyond what was expected for a subject's chronological age,
      World Health Organization
      Global status report on noncommunicable diseases 2014.
      and ultra-processed food consumption could have a role in this phenomenon.
      • Lim S.S.
      • Vos T.
      • Flaxman A.D.
      • et al.
      A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.
      Indeed, several longitudinal studies observed that ultra-processed food intake was associated with a higher incidence of dyslipidemia,
      • Rauber F.
      • Campagnolo P.D.
      • Hoffman D.J.
      • Vitolo M.R.
      Consumption of ultra-processed food products and its effects on children's lipid profiles: a longitudinal study.
      hypertension
      • Mendonca R.D.
      • Lopes A.C.
      • Pimenta A.M.
      • Gea A.
      • Martinez-Gonzalez M.A.
      • Bes-Rastrollo M.
      Ultra-Processed Food Consumption and the Incidence of Hypertension in a Mediterranean Cohort: The Seguimiento Universidad de Navarra Project.
      and cancer.
      • Fiolet T.
      • Srour B.
      • Sellem L.
      • et al.
      Consumption of ultra-processed foods and cancer risk: results from NutriNet-Sante prospective cohort.
      Moreover, both cross-sectional
      • Canella D.S.
      • Levy R.B.
      • Martins A.P.
      • et al.
      Ultra-processed food products and obesity in Brazilian households (2008-2009).
      • Monteiro C.A.
      • Moubarac J.C.
      • Levy R.B.
      • Canella D.S.
      • Louzada M.L.D.C.
      • Cannon G.
      Household availability of ultra-processed foods and obesity in nineteen European countries.
      and longitudinal studies
      • Mendonça R.D.
      • Pimenta A.M.
      • Gea A.
      • et al.
      Ultraprocessed food consumption and risk of overweight and obesity: the University of Navarra Follow-Up (SUN) cohort study.
      found a relationship with obesity.
      Very recently, a cohort study conducted in France (the NutiriNet.Santé Study) assessed the association between ultra-processed food consumption and the risk of mortality among middle-aged adults (45 years old and older).
      • Schnabel L.
      • Kesse-Guyot E.
      • Alles B.
      • et al.
      Association between ultraprocessed food consumption and risk of mortality among middle-aged adults in France [published online ahead of print February 11, 2019].
      Participants were highly selected (mainly women, and more health conscious than the general population),
      • Andreeva V.A.
      • Deschamps V.
      • Salanave B.
      • et al.
      Comparison of dietary intakes between a large online cohort study (Etude NutriNet-Sante) and a nationally representative cross-sectional study (Etude Nationale Nutrition Sante) in France: addressing the issue of generalizability in e-epidemiology.
      and all the information was based on electronic data. The authors found a positive relationship between ultra-processed food intake and the risk of mortality.
      This study, conducted on a representative sample of the noninstitutionalized adult population of Spain, aimed to elucidate the association of ultra-processed food consumption as a whole with all-cause mortality. Moreover, we examined the effect of theoretical isocaloric nonprocessed foods substitution on this relationship. Information on food groups was also provided.

      Patients and Methods

      Study Population

      Data were taken from the Study on Nutrition and Cardiovascular Risk in Spain (ENRICA), whose methods have been reported elsewhere.
      • Rodríguez-Artalejo F.
      • Graciani A.
      • Guallar-Castillón P.
      • et al.
      In brief, 12,948 individuals were selected between June 2008 and October 2010 by stratified cluster sampling to ensure that they were a representative sample of the non-institutionalized population of Spain age 18 years and older. First, the sample was stratified by province and size of the municipality. Second, clusters were selected randomly in 2 stages: municipalities and census sections. Finally, the households within each section were selected by random telephone dialing. Participants in the households were selected proportionally to the sex and age distribution of the Spanish population.
      Trained and certified personnel collected information in 3 sequential stages: (1) a telephone interview to obtain data on sociodemographic factors, health behaviors, self-rated health, and morbidity; (2) a first home visit to collect blood and urine samples, and (3) a second home visit to perform a physical examination, and to obtain habitual diet by using a computerized dietary history. More information about the sample collection process has been already published.
      • Rodríguez-Artalejo F.
      • Graciani A.
      • Guallar-Castillón P.
      • et al.
      Written informed consent was obtained from all participants. The study was approved by the Clinical Research Ethics Committees of the La Paz University Hospital in Madrid and the Hospital Clinic in Barcelona (Spain).

      Dietary Assessment and Extent and Purpose of Processing of Foods

      A validated computer-based dietary history (DH-ENRICA) was used to ascertain the participant's habitual consumption of different foods. The DH-ENRICA is a computerized questionnaire administered by a trained interviewer; it includes 880 foods that can be cooked in 29 different ways and 184 recipes for dishes commonly eaten in Spain or typical of each region. Taking into account the weekly frequency of consumption of each food and the seasonal intake, this dietary history provides an estimate in daily grams of foods that represent the average intake during the preceding year. More details about DH-ENRICA and dietary information collection have been published in Guallar-Castillon et al.
      • Guallar-Castillón P.
      • Sagardui-Villamor J.
      • Balboa-Castillo T.
      • et al.
      Validity and reproducibility of a Spanish dietary history.
      Standard food composition tables from Spain and other countries allowed calculation of the amount of energy and nutrient intake.
      • Guallar-Castillón P.
      • Sagardui-Villamor J.
      • Balboa-Castillo T.
      • et al.
      Validity and reproducibility of a Spanish dietary history.
      All recorded food items were classified according to the NOVA food classification based on the extent and purpose of industrial food processing.
      • Monteiro C.A.
      • Cannon G.
      • Moubarac J.C.
      • Levy R.B.
      • Louzada M.L.C.
      • Jaime P.C.
      The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing.
      The rationale underlying the NOVA food classification, the detailed definition of each NOVA food group, and examples of food items classified in each group have been shown elsewhere.
      Food and Agriculture Organization of the United Nations
      Guidelines on the collection of information on food processing through food consumption surveys.
      • Costa Louzada M.L.
      • Martins A.P.
      • Canella D.S.
      • et al.
      Ultra-processed foods and the nutritional dietary profile in Brazil.
      • Monteiro C.
      • Cannon G.
      • Levy R.
      • et al.
      The food system. Processing. The big issue for disease, good health, well-being.
      Foods were classified into 4 groups: (1) “unprocessed or minimally processed foods,” defined as foods that are of plant or of animal origin consumed shortly after harvesting, gathering, slaughtering, or husbanding, or foods that are altered in ways that do not add or introduce any substances, but that might involve subtracting parts of the food; (2) “processed culinary ingredients,” defined as food products extracted and refined from elements of foods, such as plant oils, animal fats, starches, and sugar, or obtained from nature (eg, salt) that are normally not consumed by themselves; (3) “processed foods” that are made by adding salt or sugar (or other substances of culinary use such as oil or vinegar) to unprocessed or minimally processed foods, in order to preserve them or to enhance their palatability; and (4) “ultra-processed foods,” which are those that were formulated mostly or entirely from substances derived from foods, with little or even no whole food content. These ingredients include modified starches, hydrogenated oils, protein isolates, and additives whose purpose is to imitate sensorial qualities of unprocessed or minimally processed foods and their culinary preparations, or to disguise undesirable qualities of the final product, such as colorants, flavorings, nonsugar sweeteners, emulsifiers, humectants, sequestrants, and firming, bulking, defoaming, anticaking, and glazing agents. The full list of the recorded foods and their NOVA classification is shown in the Supplemental Appendix (available online at http://www.mayoclinicproceedings.org). It should be noted that a food group can have foods that can be classified in different NOVA categories. For example, in the food group “yogurts and fermented milks,” natural skimmed fermented milk, natural fermented whole milk, natural skimmed yogurt, natural whole yogurt, and kefir were considered in group 1 as “unprocessed or minimally processed foods,” whereas skimmed fermented milk with fruits, whole fermented milk with fruits, skimmed flavored yogurt, skimmed yogurt with fruits, natural skimmed yogurt with sweetener, whole flavored yogurt, whole yogurt with fruits, natural whole yogurt with sweetener, whole drinking/liquid yogurt, whole drinking/liquid yogurt with fruit, natural drinking/liquid yogurt with sweetener, and natural Greek yogurt were considered in group 4 as “ultra-processed foods.”

      Mortality Ascertainment

      All-cause mortality from baseline in 2008-2010 to the end of follow-up on December 31, 2016, was obtained for 99.9% of the cohort. A computerized search was performed in the Spanish National Death Index, which contains information on the vital status of all residents in Spain. Censoring was set at the date of death or at the end of follow-up, whichever occurred first.

      Covariates

      Sex, age, educational level (no formal education, primary, and secondary or higher) were recorded. Individuals also reported whether they lived alone; whether they were current, former, or never smokers; and whether they were former drinkers. Physical activity was recorded using the European Prospective Investigation into Cancer and Nutrition cohort questionnaire. A Physical Activity Index was based on a cross-tabulation of occupational, household, and recreational activities, categorizing individuals into 4 levels of activity: inactive, moderately inactive, moderately active, and active.
      • Wareham N.J.
      • Jakes R.W.
      • Rennie K.L.
      • et al.
      Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study.
      Time spent watching television and time devoted to other sedentary activities (computer use, reading, commuting, and listening to music) were also registered in hours per week.
      • Martínez-González M.A.
      • López-Fontana C.
      • Varo J.J.
      • Sánchez-Villegas A.
      • Martinez J.A.
      Validation of the Spanish version of the physical activity questionnaire used in the Nurses' Health Study and the Health Professionals' Follow-up Study.
      Finally, the amount of medication per day and the presence of chronic conditions diagnosed by a physician (chronic respiratory disease, coronary heart disease, stroke, heart failure, osteoarthritis, cancer, and depression requiring treatment) were also self-reported.

      Statistical Analysis

      Of the 12,948 study participants, 1050 of them were excluded because of inconsistent dietary data (total daily energy intake outside the range of 800-5000 kcal in men, or 500-4000 in women; n=60), missing information on diet (n=887), or covariates (n=103). Thus, the analyses were conducted with 11,898 individuals (5890 men and 6008 women).
      For each participant, the proportion of total energy from ultra-processed foods was calculated, and the total sample was divided into quartiles (cutoff points were set at 14.1%, 23.0%, and 33.1%). Cox proportional hazards models were used to assess the association between quartiles of ultra-processed food intake (in percentage of total energy intake) and mortality. The first quartile (lowest intake) was used as the reference. Analyses were weighted to account for the complex sampling design, and the variances were also corrected to calculate appropriate 95% CIs. Linear trend was calculated using quartiles of ultra-processed food as a continuous variable. Schoenfeld residuals were plotted against time to detect violations of the proportional hazard assumption. Finally, ultra-proceed foods were also expressed in grams per day per kilogram weight of the participant. Bodyweight was assessed using electronic scales (model Seca 841; Seca Deutschland, Hamburg, Germany; precision to 0.1 kg).
      Cox models were built with 4 successive levels of additional adjustments: model 1 was adjusted for age and sex; model 2 was adjusted for socioeconomic factors (educational level and living alone); model 3 was adjusted for lifestyle factors (smoking status, former drinker, physical activity, time watching television, and time devoted to other sedentary activities), and model 4 was adjusted for clinical factors (number of medications per day and presence of chronic conditions). We also performed several sensitivity analyses.
      We built isocaloric substitution models that simultaneously included energy intake, the percentages of energy derived from ultra-processed foods, from processed culinary ingredients, and processed food or unprocessed or minimally processed foods, as appropriate, and other potential confounding variables. Nonlinear relations were identified using restricted cubic spline analyses with 3 knots or inflection points placed at the 10th, 50th, and 90th percentiles. The coefficients from these models can be interpreted as the estimated association of substituting a certain percentage of energy from ultra-processed foods with the equivalent energy from processed foods or unprocessed or minimally processed foods while holding constant the intake of total energy and the energy from the corresponding nonreplaced NOVA groups.
      • Hu F.B.
      • Stampfer M.J.
      • Rimm E.
      • et al.
      Dietary fat and coronary heart disease: a comparison of approaches for adjusting for total energy intake and modeling repeated dietary measurements.
      Secondary analyses were performed for nutrient intake from ultra-processed foods (as a percentage of total energy or as caloric density), considering the main food groups contributing to ultra-processed food intake. These variables were divided into quartiles, tertiles, or median according to the number of participants with intake in the corresponding categories and always using the lowest category as a reference. Analyses were performed with Stata software version 14 for Windows (StataCorp LP), and statistical significance was set at P<.05. The survey command was used in the analyses to account for the complex sampling.

      Results

      The cohort comprised 11,898 subjects (mean age, 46.9 years; 50.5% female) who had an average consumption of ultra-processed food of 385 g/d, corresponding to 24.4% of the total energy intake. Those in the first quartile of ultra-processed foods consumed a mean of 8.7% (mean, 156 g), versus 42.8% (mean, 641 g) consumed by those in the highest quartile (Supplemental Figure 1, Supplemental Table 1, available online at http://www.mayoclinicproceedings.org). After a mean of 7.7 years and 93,599 person-years followed up, 440 deaths occurred.
      Subjects that consumed more ultra-processed foods also ingested more energy, were younger, were more frequently current smokers, were more educated, had a higher activity index, watched fewer hours of television but dedicated more time to other sedentary activities, took less medication per day, and had lower prevalence of stroke, heart failure, osteoarthritis, and cancer but higher rates of depression (Table 1). They also had a higher intake of carbohydrates, simple sugars, total fat, saturated fatty acids, polyunsaturated fatty acids, trans fatty acids, and sodium (Supplemental Table 1, Supplemental Table 1, available online at http://www.mayoclinicproceedings.org). The main food groups contributing to ultra-processed food intake were meat and meat products (17.1%), cakes and pastries (13.6%), cookies (9.2%), yogurts and fermented milk (8.8%), jams and confectionery (7.4%), and precooked dishes (7.1%). These food groups had a different relative contribution among ultra-processed quartiles (Supplemental Table 2).
      Table 1Baseline Characteristics of the Cohort Participants According to Quartiles of Ultra-processed Food Consumption (% of Energy) in the ENRICA Study (2008-2010)
      CharacteristicsAll subjects (N=11,898)Quartiles of ultra-processed food consumption (% total energy)
      Interquartiles ranges are: 0%-14.08% for Q1; 14.09%-23.00% for Q2; 23.01%-33.14% for Q3; and 33.15%-100% for Q4.
      Q1, lowest (n=2976)Q2 (n=2974)Q3 (n=2974)Q4, highest (n=2974)P for linear trend
      Total energy (kcal/d), mean ± SE2173.9 ± 9.01970.0 ± 13.32078.0 ± 13.42263.5 ± 17.42378.9 ± 15.8<.001
      Ultra-processed food consumption (% of energy), mean ± SE24.47 ± 0.178.68 ± 0.0818.60 ± 0.0527.82 ± 0.0642.83 ± 0.19<.001
      Ultra-processed food consumption (g/d), mean ± SE384.70 ± 4.30155.50 ± 2.90304.70 ± 5.10436.90 ± 6.20641.20 ± 9.00<.001
      Weight (kg), mean ± SE73.80 ± 0.2074.60 ± 0.3074.10 ± 0.4073.60 ± 0.3073.00 ± 0.40<.001
      Ultra-processed food consumption (grams per d/weight), mean ± SE5.34 ± 0.062.12 ± 0.044.18 ± 0.076.08 ± 0.099.01 ± 0.13<.001
      Women (%)50.5%49.1%51.8%50.1%51.0%.39
      Age (y), mean ± SE46.9 ± 0.2754.93 ± 0.4049.40 ± 0.3844.60 ± 0.4438.79 ± 0.40<.001
      Educational level, (%)<.001
       No formal education30.0%40.5%31.2%26.1%22.1%
       Primary42.0%32.6%40.5%45.1%49.7%
       Secondary or higher28.1%26.9%28.4%28.8%28.2%
      Living alone (%)7.8%9.0%8.8%7.1%6.2%<.001
      Smoking status (%)<.001
       Current smoker27.5%24.7%27.0%28.4%30.0%
       Former smoker24.7%30.8%25.7%22.9%19.6%
       Never smoker47.8%44.6%47.3%48.7%50.4%
      Former drinker (%)5.5%5.3%5.5%5.8%5.4%.93
      Physical activity index (%)<.001
       Inactive27.7%33.2%28.6%25.6%23.4%
       Moderately inactive33.8%35.2%33.3%33.3%33.3%
       Moderately active23.0%20.1%23.7%24.2%23.9%
       Active15.6%11.6%14.4%17.0%19.4%
      Time watching television (h/wk), mean ± SE13.67 ± 0.1214.43 ± 0.2413.85 ± 0.2313.35 ± 0.2113.16 ± 0.23<.001
      Time devoted to other sedentary activities (h/wk), mean ± SE15.52 ± 0.1613.69 ± 0.2915.11 ± 0.2616.24 ± 0.3317.03 ± 0.28<.001
      Number of medications per day, mean ± SE0.84 ± 0.021.18 ± 0.040.95 ± 0.040.71 ± 0.030.53 ± 0.03<.001
      Chronic conditions (%)
       Chronic respiratory disease6.4%6.3%6.8%6.5%5.9%.72
       Coronary heart disease0.6%0.9%0.6%0.5%0.4%.15
       Stroke0.4%0.7%0.6%0.2%0.3%.03
       Heart failure1.5%1.7%1.9%1.1%1.1%.03
       Osteoarthritis20.7%28.2%24.6%17.2%12.9%<.001
       Cancer1.0%1.6%0.9%0.9%0.7%.01
       Depression6.6%6.4%7.3%5.4%7.5%.02
      a Interquartiles ranges are: 0%-14.08% for Q1; 14.09%-23.00% for Q2; 23.01%-33.14% for Q3; and 33.15%-100% for Q4.
      In all models, participants in the highest quartile of ultra-processed food intake had higher mortality risk when compared with those in the lowest quartile; in model 4, the hazard ratio (HR) for mortality was 1.44 (95% CI, 1.01-2.07; P for trend=.03). The corresponding mortality risk when ultra-processed food consumption was expressed in grams per day per kilogram weight was 1.46 (95% CI, 1.04-2.05; P for trend=.03; Table 2). Sensitivity analyses provided similar results (Supplemental Table 3, Supplemental Table 1, available online at http://www.mayoclinicproceedings.org).
      Table 2Mortality risk according to ultra-processed food consumption quartiles in the ENRICA Study (N=11,898)
      HR = hazard ratio.
      Q1 (lowest)Q2Q3Q4 (highest)P for linear trend
      Quartiles of ultra-processed food consumption (% energy)
       n2976297429742974
       Deaths15810510374
       Person-years23,30823,37823,43823,475
       Model 1, HR (95% CI)
      HR = hazard ratio.
      1 (Reference)0.97 (0.73-1.30)1.28 (0.95-1.73)1.44 (1.00-2.06).02
       Model 2, HR (95% CI)1 (Reference)0.98 (0.73-1.31)1.30 (0.96-1.74)1.51 (1.05-2.16).01
       Model 3, HR (95% CI)1 (Reference)0.94 (0.70-1.26)1.25 (0.93-1.68)1.42 (0.99-2.04).03
       Model 4, HR (95% CI)1 (Reference)0.91 (0.67-1.23)1.23 (0.91-1.67)1.44 (1.01-2.07).03
      Quartiles of ultra-processed food consumption (grams per d/weight)
       n2951295029502950
       Deaths1481259859
       Person-years23,18123,18323,22223,281
       Model 1, HR (95% CI)
      Model 1: adjusted for sex and age. Model 2: adjusted as in model 1 plus, educational level, and living alone. Model 3: adjusted as in model 2 plus, smoking status, former drinker, physical activity index, time watching television, and time devoted to other sedentary activities. Model 4: adjusted as in model 3 plus the number of medications per day, and specific chronic conditions diagnosed by a physician (chronic respiratory disease, coronary heart disease, stroke, heart failure, osteoarthritis, cancer, and depression).
      1 (Reference)1.07 (0.81-1.41)1.21 (0.91-1.62)1.48 (1.05-2.10).02
       Model 2, HR (95% CI)1 (Reference)1.08 (0.82-1.43)1.23 (0.93-1.65)1.39 (0.97-1.99).01
       Model 3, HR (95% CI)1 (Reference)1.08 (0.82-1.43)1.19 (0.89-1.60)1.48 (1.06-2.07).02
       Model 4, HR (95% CI)1 (Reference)1.08 (0.81-1.44)1.21 (0.89-1.64)1.46 (1.04-2.05).03
      a HR = hazard ratio.
      b Model 1: adjusted for sex and age. Model 2: adjusted as in model 1 plus, educational level, and living alone. Model 3: adjusted as in model 2 plus, smoking status, former drinker, physical activity index, time watching television, and time devoted to other sedentary activities. Model 4: adjusted as in model 3 plus the number of medications per day, and specific chronic conditions diagnosed by a physician (chronic respiratory disease, coronary heart disease, stroke, heart failure, osteoarthritis, cancer, and depression).
      In model 4, the hazard ratio of the iso-caloric substitution of ultra-processed foods with processed foods was less than 1 but did not achieve statistical significance (Figure A ). However, when replaced with unprocessed or minimally processed foods, an inverse and nonlinear dose-response was observed (P for nonlinearity=.04), starting to be significant when replacing 8.2% of ultra-processed food consumption (Figure B).
      Figure thumbnail gr1
      Figure(A) Risk of death associated with isocaloric replacement of ultra-processed food (in percent of energy) with processed foods (in percent of energy). The y-axis shows the predicted hazard ratios (HRs) for total mortality, and the x-axis shows the percentage of energy from nonprocessed food intake as a continuous variable. Lines are restricted cubic splines, showing the shape of the isocaloric substitution. The black line represents HR, and the dashed lines indicate the upper and lower 95% CIs. The knots were located at the 10th, 50th, and 90th percentiles (2.28%, 6.03%, and 11.07% of energy intake, respectively). Adjusted as in model 4. (P nonlinearity=.53). (B) Risk of death associated with isocaloric replacement of ultra-processed food (in percent of energy) with unprocessed or minimally processed foods (in percent of energy). The y-axis shows the predicted HRs for total mortality, and the x-axis shows the percentage of energy from nonprocessed food intake as a continuous variable. Lines are restricted cubic splines, showing the shape of the isocaloric substitution. The black line plots the HR, and the dashed lines indicate the upper and lower 95% CIs. The knots were located at the 10th, 50th and 90th percentiles (4.52%, 6.89%, and 10.3% of energy intake, respectively). Adjusted as in model 4. (P nonlinearity=.04).
      When comparing the highest versus the lowest quartile of nutrients intake from ultra-processed foods, consumption of total carbohydrates, simple sugars, saturated fatty acids, and trans fatty acids contributed the most to increased mortality risk, although only trans fatty acids reached statistical significance (HR, 1.39; 95% CI, 1.00-1.92; P for trend=.05; Supplemental Figure 2, Supplemental Table 1, available online at http://www.mayoclinicproceedings.org). When analyzing ultra-processed food consumption by food groups, the intake of yogurts and fermented milks (those classified as ultra-processed), cakes and pastries, and cookies contributed the most to an increased mortality risk, but it only reached statistical significance for yogurts and fermented milks (HR, 1.37; 95% CI, 1.02-1.86) with a nonsignificant trend (P for linear trend=.09). The other groups of ultra-processed foods (breads, breakfast cereals, dairy desserts, meat and meat products, jams and confectionery, sauces and dressings, soft drinks, packaged fruit juices, and nectars) were not associated with increased risk of mortality. (Supplemental Figure 3, Supplemental Table 1, available online at http://www.mayoclinicproceedings.org).

      Discussion

      Principal Findings

      In this prospective study, based on a representative sample of the noninstitutionalized adult population of Spain, an increased intake of ultra-processed food was associated with higher mortality after 7.7 years of follow-up. Participants in which ultra-processed food intake contributed more than 33% of total energy intake had a 44% higher all-cause mortality risk compared with those in whom ultra-processed food intake contributed up to 14% of total energy intake.

      Interpretation and Comparison with Other Studies

      To our knowledge, this is the first prospective epidemiologic study performed in a large and representative national cohort that evaluated the association between ultra-processed food consumption (as a whole) and risk for all-cause mortality in participants age 18 years and older.
      Our results are in line with those obtained in the NutriNet-Santé Study.
      • Schnabel L.
      • Kesse-Guyot E.
      • Alles B.
      • et al.
      Association between ultraprocessed food consumption and risk of mortality among middle-aged adults in France [published online ahead of print February 11, 2019].
      According to our results, the strength of the association is probably higher than that presented in the NutriNet-Santé Study. This is especially important because young generations are increasing their consumption of ultra-processed food, and its consequences will be seen in the future.
      In Spain, consumption of ultra-processed food intake was 24.4% of total dietary calories, being consistent with previous data from the Data Food Networking Database databank showing that 20.3% of total purchased dietary energy came from ultra-processed foods.
      • Monteiro C.A.
      • Moubarac J.C.
      • Levy R.B.
      • Canella D.S.
      • Louzada M.L.D.C.
      • Cannon G.
      Household availability of ultra-processed foods and obesity in nineteen European countries.
      Spain is a country with low ultra-processed food consumption when compared with other Western countries, such as Canada (61.7%),
      • Moubarac J.C.
      • Batal M.
      • Martins A.P.
      • et al.
      Processed and ultra-processed food products: consumption trends in Canada from 1938 to 2011.
      the United States (57.9%),
      • Baraldi L.G.
      • Martinez Steele E.
      • Canella D.S.
      • Monteiro C.A.
      Consumption of ultra-processed foods and associated sociodemographic factors in the USA between 2007 and 2012: Evidence from a nationally representative cross-sectional study.
      the United Kingdom (53%),
      • Adams J.
      • White M.
      Characterisation of UK diets according to degree of food processing and associations with socio-demographics and obesity: cross-sectional analysis of UK National Diet and Nutrition Survey (2008-12).
      and France (35.9%),
      • Julia C.
      • Martinez L.
      • Allès B.
      • et al.
      Contribution of ultra-processed foods in the diet of adults from the French NutriNet-Santé study.
      or developing countries such as Brazil (29.6%).
      • Costa Louzada M.L.
      • Martins A.P.
      • Canella D.S.
      • et al.
      Ultra-processed foods and the nutritional dietary profile in Brazil.
      This could be explained as cooking at home being part of the Mediterranean diet, which is also rich in unprocessed or minimally processed foods.
      • Sofi F.
      • Macchi C.
      • Abbate R.
      • Gensini G.F.
      • Casini A.
      Mediterranean diet and health status: an updated meta-analysis and a proposal for a literature-based adherence score.
      United Nations Educational, Scientific and Cultural Organization
      Mediterranean Diet: Cyprus, Croatia, Spain, Greece, Italy, Morocco and Portugal Inscribed in 2013 (8.COM) on the Representative List of the Intangible Cultural Heritage of Humanity.
      • Willett W.C.
      • Sacks F.
      • Trichopoulou A.
      • et al.
      Mediterranean diet pyramid: a cultural model for healthy eating.
      However, it is also known that the Spanish population has been drifting away from this traditional pattern to adopt a less healthy diet,
      • León-Muñoz L.M.
      • Guallar-Castillón P.
      • Graciani A.
      • et al.
      Adherence to the Mediterranean diet pattern has declined in Spanish adults.
      especially among young people,
      • Garcia-Meseguer M.J.
      • Burriel F.C.
      • Garcia C.V.
      • Serrano-Urrea R.
      Adherence to Mediterranean diet in a Spanish university population.
      supporting the estimates that ultra-processed foods consumption will continue to increase.
      • Kelly B.
      • Jacoby E.
      Public Health Nutrition special issue on ultra-processed foods.
      The association of some groups of ultra-processed foods with mortality has been studied previously. In the United States, positive associations were found between fast food and sugar-sweetened beverage consumption and mortality.
      • Barrington W.E.
      • White E.
      Mortality outcomes associated with intake of fast-food items and sugar-sweetened drinks among older adults in the Vitamins and Lifestyle (VITAL) study.
      In addition, a recent meta-analysis showed a nonlinear 7% higher risk with an increased intake of sugar-sweetened beverages up to 250 mL/d.
      • Garcia-Meseguer M.J.
      • Burriel F.C.
      • Garcia C.V.
      • Serrano-Urrea R.
      Adherence to Mediterranean diet in a Spanish university population.
      An increased risk of mortality has also been shown with meats and processed meats.
      • Pan A.
      • Sun Q.
      • Bernstein A.M.
      • et al.
      Red meat consumption and mortality: results from 2 prospective cohort studies.
      • Rohrmann S.
      • Overvad K.
      • Bueno-de-Mesquita H.B.
      • et al.
      Meat consumption and mortality--results from the European Prospective Investigation into Cancer and Nutrition.
      • Schwingshackl L.
      • Schwedhelm C.
      • Hoffmann G.
      • et al.
      Food groups and risk of all-cause mortality: a systematic review and meta-analysis of prospective studies.
      Regarding dairy products, studies showed controversial results. Although a recent meta-analysis of 29 prospective cohort studies demonstrated neutral associations between milk and dairy products and mortality; there was not a distinction between processed and ultra-processed foods.
      • Guo J.
      • Astrup A.
      • Lovegrove J.A.
      • Gijsbers L.
      • Givens D.I.
      • Soedamah-Muthu S.S.
      Milk and dairy consumption and risk of cardiovascular diseases and all-cause mortality: dose-response meta-analysis of prospective cohort studies.
      In the present analysis, sugared, sweetened, flavored, and additive-added dairy products were considered as ultra-processed. On the contrary, milk, plain yogurts, and fresh and cured cheeses were not included in this category. Our results suggest that the intake of ultra-processed yogurts and fermented milks is associated with increased mortality, reinforcing the idea of considering the extent of processing when studying dairy product. Moreover, the obtained isocaloric replacement results are supported by different studies that used national household data to estimate the contribution of dietary trends for risk of cardiovascular disease and all-cause mortality. In the United Kingdom, a reduction of 13% of cardiovascular disease mortality was projected by 2030 if dietary intake of ultra-processed and processed foods were entirely replaced with unprocessed or minimally processed foods.
      • Moreira P.V.
      • Baraldi L.G.
      • Moubarac J.C.
      • et al.
      Comparing different policy scenarios to reduce the consumption of ultra-processed foods in UK: impact on cardiovascular disease mortality using a modelling approach.
      In Brazil, using a similar approach, an 11% cardiovascular mortality reduction was estimated if ultra-processed foods were reduced by 50% and substituted with unprocessed or minimally processed foods, plus an additional 50% reduction in processed culinary ingredients.
      • Moreira P.V.
      • Hyseni L.
      • Moubarac J.C.
      • et al.
      Effects of reducing processed culinary ingredients and ultra-processed foods in the Brazilian diet: a cardiovascular modelling study.
      How ultra-processed food consumption increases the risk of all-cause mortality could depend on a number of factors. Ultra-processed foods have a high-energy density that is less satiating, highly accessible
      • Monteiro C.A.
      • Cannon G.
      • Moubarac J.C.
      • Levy R.B.
      • Louzada M.L.C.
      • Jaime P.C.
      The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing.
      • Monteiro C.
      • Cannon G.
      • Levy R.
      • et al.
      The food system. Processing. The big issue for disease, good health, well-being.
      • Fardet A.
      Minimally processed foods are more satiating and less hyperglycemic than ultra-processed foods: a preliminary study with 98 ready-to-eat foods.
      and prone to causing inadvertent overconsumption, which has also been associated with mortality.
      • Nagai M.
      • Ohkubo T.
      • Miura K.
      • et al.
      Association of Total Energy Intake with 29-Year Mortality in the Japanese: NIPPON DATA80.
      Moreover, ultra-processed foods are industrial formulations containing high quantities of saturated fatty acid, trans fatty acids, hydrogenated oils, starches, free sugars, and salt, plus food additives used to imitate the sensory qualities of natural foods, or to disguise undesirable qualities of the final product, such as colorants, flavorings, artificial sweeteners, and emulsifiers.
      • Luiten C.M.
      • Steenhuis I.H.
      • Eyles H.
      • Ni Mhurchu C.
      • Waterlander W.E.
      Ultra-processed foods have the worst nutrient profile, yet they are the most available packaged products in a sample of New Zealand supermarkets.
      • Scrinis G.
      • Monteiro C.A.
      Ultra-processed foods and the limits of product reformulation.
      Harmful effects of some of these nutritional compounds have been studied widely. Trans fatty acid intake has been related to an increase in all-cause mortality risk in several populations,
      • Chien K.L.
      • Lin H.J.
      • Hsu H.C.
      • et al.
      Comparison of predictive performance of various fatty acids for the risk of cardiovascular disease events and all-cause deaths in a community-based cohort.
      • Kiage J.N.
      • Merrill P.D.
      • Robinson C.J.
      • et al.
      Intake of trans fat and all-cause mortality in the Reasons for Geographical and Racial Differences in Stroke (REGARDS) cohort.
      as has total sugar intake.
      • Tasevska N.
      • Park Y.
      • Jiao L.
      • Hollenbeck A.
      • Subar A.F.
      • Potischman N.
      Sugars and risk of mortality in the NIH-AARP Diet and Health Study.
      The association of saturated fatty acid intake and mortality is controversial,
      • de Souza R.J.
      • Mente A.
      • Maroleanu A.
      • et al.
      Intake of saturated and trans unsaturated fatty acids and risk of all cause mortality, cardiovascular disease, and type 2 diabetes: systematic review and meta-analysis of observational studies.
      • Wakai K.
      • Naito M.
      • Date C.
      • Iso H.
      • Tamakoshi A.
      JACC Study Group
      Dietary intakes of fat and total mortality among Japanese populations with a low fat intake: the Japan Collaborative Cohort (JACC) Study.
      highlighting the need to take into account the food sources of saturated fatty acids.
      • Guasch-Ferré M.
      • Babio N.
      • Martínez-González M.A.
      • et al.
      Dietary fat intake and risk of cardiovascular disease and all-cause mortality in a population at high risk of cardiovascular disease.
      In addition, a meta-analysis of 23 cohort studies found a U-shaped association between sodium intake and mortality.
      • Graudal N.
      • Jürgens G.
      • Baslund B.
      • Alderman M.H.
      Compared with usual sodium intake, low- and excessive-sodium diets are associated with increased mortality: a meta-analysis.
      In our cohort, the separate analysis of these nutrients from ultra-processed foods increase the risk of all-cause mortality, although only trans fatty acids reached a significant association. These results support the application of food processing classification beyond conventional food classifications and the report on single nutrients
      • Luiten C.M.
      • Steenhuis I.H.
      • Eyles H.
      • Ni Mhurchu C.
      • Waterlander W.E.
      Ultra-processed foods have the worst nutrient profile, yet they are the most available packaged products in a sample of New Zealand supermarkets.
      ; because the concept of ultra-processed foods does not just focus attention on these “traditional” harmful nutrients but to a wider range of food additives contained in this type of foods. Although food additive safety is rigorously controlled by the European authorities,
      • Carocho M.
      • Barreiro M.F.
      • Morales P.
      • Ferreira I.C.F.R.
      Adding molecules to food, pros and cons: a review on synthetic and natural food additives.
      the detrimental health effects in the long term and the potential interaction consequences between the different compounds are largely unknown.
      • Scrinis G.
      • Monteiro C.A.
      Ultra-processed foods and the limits of product reformulation.
      Finally, consumption of ultra-processed foods was also associated with a poorer diet quality (with less consumption of fiber, fruit, and vegetables)
      • Martínez Steele E.
      • Popkin B.M.
      • Swinburn B.
      • Monteiro C.A.
      The share of ultra-processed foods and the overall nutritional quality of diets in the US: evidence from a nationally representative cross-sectional study.
      • Moubarac J.C.
      • Batal M.
      • Louzada M.L.
      • Martinez Steele E.
      • Monteiro C.A.
      Consumption of ultra-processed foods predicts diet quality in Canada.
      and unhealthy lifestyle (smoking and sedentary behaviors),
      • Julia C.
      • Martinez L.
      • Allès B.
      • et al.
      Contribution of ultra-processed foods in the diet of adults from the French NutriNet-Santé study.
      • Costa C.D.S.
      • Flores T.R.
      • Wendt A.
      • Neves R.G.
      • Assunção M.C.F.
      • Santos I.S.
      Sedentary behavior and consumption of ultra-processed foods by Brazilian adolescents: Brazilian National School Health Survey (PeNSE), 2015.
      all contributing to increased mortality.
      • Moodie R.
      • Stuckler D.
      • Monteiro C.
      • et al.
      Lancet NCD Action Group
      Profits and pandemics: prevention of harmful effects of tobacco, alcohol, and ultra-processed food and drink industries.
      • Schwingshackl L.
      • Schwedhelm C.
      • Hoffmann G.
      • et al.
      Food groups and risk of all-cause mortality: a systematic review and meta-analysis of prospective studies.
      • Martínez-Gómez D.
      • Guallar-Castillón P.
      • León-Muñoz L.M.
      • López-García E.
      • Rodríguez-Artalejo F.
      Combined impact of traditional and non-traditional health behaviors on mortality: a national prospective cohort study in Spanish older adults.
      However, further mechanistic studies to know how and to which extent ultra-processed foods could affect health should be performed.

      Strengths and Limitations of the Study

      This study has some strengths, including its prospective design and a relatively large sample size representative of the adult Spanish population, which broadens the generalization of the results. Food consumption was collected in detail with a validated dietary history including a wide variety of foods and their cooking methods.
      • Guallar-Castillón P.
      • Sagardui-Villamor J.
      • Balboa-Castillo T.
      • et al.
      Validity and reproducibility of a Spanish dietary history.
      Finally, many confounding factors were considered in the analysis; however, there are some limitations. First, dietary information was obtained only at baseline, assuming no time changes in dietary intake and probably underestimating the real impact of ultra-processed foods on mortality. Second, although there was a consensus between authors about NOVA classification of the foods obtained by the dietary history, some degree of misclassification among ultra-processed food categories cannot be ruled out. In this regard, we were aware that NOVA classification has been criticized (mainly because of the broad definition of ultra-processed food),
      • Gibney M.J.
      • Forde C.G.
      • Mullally D.
      • Gibney E.R.
      Ultra-processed foods in human health: a critical appraisal.
      but it is the most used and recommended for classifying ultra-processed foods in public health nutrition.
      • Kelly B.
      • Jacoby E.
      Public Health Nutrition special issue on ultra-processed foods.
      Third, the relatively low number of deaths precluded meaningful analyses by specific causes of death. Last, a certain degree of residual confounding can still exist, and nondifferential misclassification affecting both, exposure and disease, can still operate, leading to a dilution bias of the true effect.

      Conclusions and Policy Implications

      In conclusion, an increased intake of ultra-processed food was associated with a higher risk of mortality. Moreover, the theoretical isocaloric substitution of ultra-processed food by unprocessed or minimally processed foods would suppose a reduction of the mortality risk. These results need to be confirmed by other large-scale, population-based, prospective studies in different countries and settings. Further study is also needed to identify the specific foods that mostly account for this association and to investigate possible relationships with specific causes of death. However, our findings support the epidemiologic evidence about the harmful effects of ultra-processed foods, and thus, the necessity to implement actions, such as the development of new nutritional policies and guides, both for the population and the industry. Ultra-processed food intake can be considered one of the major challenges that governments will have to face in the next decades, and one of the main opportunities for nutritional prevention.

      Supplemental Online Material

      Supplemental material can be found online at http://www.mayoclinicproceedings.org. Supplemental material attached to journal articles has not been edited, and the authors take responsibility for the accuracy of all data.

      References

        • Ludwig D.S.
        Technology, diet, and the burden of chronic disease.
        JAMA. 2011; 305: 1352-1353
        • Wrangham R.
        The evolution of human nutrition.
        Curr Biol. 2013; 23: R354-355
        • Monteiro C.A.
        • Moubarac J.C.
        • Cannon G.
        • Ng S.W.
        • Popkin B.
        Ultra-processed products are becoming dominant in the global food system.
        Obes Rev. 2013; 14: 21-28
        • Food and Agriculture Organization of the United Nations
        Guidelines on the collection of information on food processing through food consumption surveys.
        Food and Agriculture Organization of the United Nations, Rome2015
        • Luiten C.M.
        • Steenhuis I.H.
        • Eyles H.
        • Ni Mhurchu C.
        • Waterlander W.E.
        Ultra-processed foods have the worst nutrient profile, yet they are the most available packaged products in a sample of New Zealand supermarkets.
        Public Health Nutr. 2016; 19: 530-538
        • Monteiro C.A.
        • Cannon G.
        • Moubarac J.C.
        • Levy R.B.
        • Louzada M.L.C.
        • Jaime P.C.
        The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing.
        Public Health Nutr. 2018; 21: 5-17
        • Rauber F.
        • da Costa Louzada M.L.
        • Steele E.M.
        • Millett C.
        • Monteiro C.A.
        • Levy R.B.
        Ultra-processed food consumption and chronic non-communicable diseases-related dietary nutrient profile in the UK (2008-2014).
        Nutrients. 2018; 10
        • Martínez Steele E.
        • Monteiro C.A.
        Association between dietary share of ultra-processed foods and urinary concentrations of phytoestrogens in the US.
        Nutrients. 2017; 9
        • Scrinis G.
        • Monteiro C.A.
        Ultra-processed foods and the limits of product reformulation.
        Public Health Nutr. 2018; 21: 247-252
        • Martínez Steele E.
        • Popkin B.M.
        • Swinburn B.
        • Monteiro C.A.
        The share of ultra-processed foods and the overall nutritional quality of diets in the US: evidence from a nationally representative cross-sectional study.
        Popul Health Metr. 2017; 15: 6
        • Moubarac J.C.
        • Batal M.
        • Louzada M.L.
        • Martinez Steele E.
        • Monteiro C.A.
        Consumption of ultra-processed foods predicts diet quality in Canada.
        Appetite. 2017; 108: 512-520
        • Monteiro C.A.
        Nutrition and health. The issue is not food, nor nutrients, so much as processing.
        Public Health Nutr. 2009; 12: 729-731
        • Moodie R.
        • Stuckler D.
        • Monteiro C.
        • et al.
        • Lancet NCD Action Group
        Profits and pandemics: prevention of harmful effects of tobacco, alcohol, and ultra-processed food and drink industries.
        Lancet. 2013; 381: 670-679
        • Baraldi L.G.
        • Martinez Steele E.
        • Canella D.S.
        • Monteiro C.A.
        Consumption of ultra-processed foods and associated sociodemographic factors in the USA between 2007 and 2012: Evidence from a nationally representative cross-sectional study.
        BMJ Open. 2018; 8: e020574
        • Moubarac J.C.
        • Batal M.
        • Martins A.P.
        • et al.
        Processed and ultra-processed food products: consumption trends in Canada from 1938 to 2011.
        Can J Diet Pract Res. 2014; 75: 15-21
        • Julia C.
        • Martinez L.
        • Allès B.
        • et al.
        Contribution of ultra-processed foods in the diet of adults from the French NutriNet-Santé study.
        Public Health Nutr. 2018; 21: 27-37
        • Slimani N.
        • Deharveng G.
        • Southgate D.A.
        • et al.
        Contribution of highly industrially processed foods to the nutrient intakes and patterns of middle-aged populations in the European Prospective Investigation into Cancer and Nutrition study.
        Eur J Clin Nutr. 2009; 63: S206-225
        • Costa Louzada M.L.
        • Martins A.P.
        • Canella D.S.
        • et al.
        Ultra-processed foods and the nutritional dietary profile in Brazil.
        Rev Saude Publica. 2015; 49: 38
        • Cediel G.
        • Reyes M.
        • da Costa Louzada M.L.
        • et al.
        Ultra-processed foods and added sugars in the Chilean diet (2010).
        Public Health Nutr. 2018; 21: 125-133
        • Kelly B.
        • Jacoby E.
        Public Health Nutrition special issue on ultra-processed foods.
        Public Health Nutr. 2018; 21: 1-4
        • World Health Organization
        Global status report on noncommunicable diseases 2014.
        World Health Organization, Geneva2014
        • Lim S.S.
        • Vos T.
        • Flaxman A.D.
        • et al.
        A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.
        Lancet. 2012; 380: 2224-2260
        • Rauber F.
        • Campagnolo P.D.
        • Hoffman D.J.
        • Vitolo M.R.
        Consumption of ultra-processed food products and its effects on children's lipid profiles: a longitudinal study.
        Nutr Metab Cardiovasc Dis. 2015; 25: 116-122
        • Mendonca R.D.
        • Lopes A.C.
        • Pimenta A.M.
        • Gea A.
        • Martinez-Gonzalez M.A.
        • Bes-Rastrollo M.
        Ultra-Processed Food Consumption and the Incidence of Hypertension in a Mediterranean Cohort: The Seguimiento Universidad de Navarra Project.
        Am J Hypertens. 2017; 30: 358-366
        • Fiolet T.
        • Srour B.
        • Sellem L.
        • et al.
        Consumption of ultra-processed foods and cancer risk: results from NutriNet-Sante prospective cohort.
        BMJ. 2018; 360: k322
        • Canella D.S.
        • Levy R.B.
        • Martins A.P.
        • et al.
        Ultra-processed food products and obesity in Brazilian households (2008-2009).
        PLoS One. 2014; 9: e92752
        • Monteiro C.A.
        • Moubarac J.C.
        • Levy R.B.
        • Canella D.S.
        • Louzada M.L.D.C.
        • Cannon G.
        Household availability of ultra-processed foods and obesity in nineteen European countries.
        Public Health Nutr. 2018; 21: 18-26
        • Mendonça R.D.
        • Pimenta A.M.
        • Gea A.
        • et al.
        Ultraprocessed food consumption and risk of overweight and obesity: the University of Navarra Follow-Up (SUN) cohort study.
        Am J Clin Nutr. 2016; 104: 1433-1440
        • Schnabel L.
        • Kesse-Guyot E.
        • Alles B.
        • et al.
        Association between ultraprocessed food consumption and risk of mortality among middle-aged adults in France [published online ahead of print February 11, 2019].
        (JAMA Intern Med)
        • Andreeva V.A.
        • Deschamps V.
        • Salanave B.
        • et al.
        Comparison of dietary intakes between a large online cohort study (Etude NutriNet-Sante) and a nationally representative cross-sectional study (Etude Nationale Nutrition Sante) in France: addressing the issue of generalizability in e-epidemiology.
        Am J Epidemiol. 2016; 184: 660-669
        • Rodríguez-Artalejo F.
        • Graciani A.
        • Guallar-Castillón P.
        • et al.
        Rev Esp Cardiol. 2011; 64: 876-882
        • Guallar-Castillón P.
        • Sagardui-Villamor J.
        • Balboa-Castillo T.
        • et al.
        Validity and reproducibility of a Spanish dietary history.
        PLoS One. 2014; 9: e86074
        • Monteiro C.
        • Cannon G.
        • Levy R.
        • et al.
        The food system. Processing. The big issue for disease, good health, well-being.
        World Nutr. 2012; 3: 527-569
        • Wareham N.J.
        • Jakes R.W.
        • Rennie K.L.
        • et al.
        Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study.
        Public Health Nutr. 2003; 6: 407-413
        • Martínez-González M.A.
        • López-Fontana C.
        • Varo J.J.
        • Sánchez-Villegas A.
        • Martinez J.A.
        Validation of the Spanish version of the physical activity questionnaire used in the Nurses' Health Study and the Health Professionals' Follow-up Study.
        Public Health Nutr. 2005; 8: 920-927
        • Hu F.B.
        • Stampfer M.J.
        • Rimm E.
        • et al.
        Dietary fat and coronary heart disease: a comparison of approaches for adjusting for total energy intake and modeling repeated dietary measurements.
        Am J Epidemiol. 1999; 149: 531-540
        • Adams J.
        • White M.
        Characterisation of UK diets according to degree of food processing and associations with socio-demographics and obesity: cross-sectional analysis of UK National Diet and Nutrition Survey (2008-12).
        Int J Behav Nutr Phys Act. 2015; 12: 160
        • Sofi F.
        • Macchi C.
        • Abbate R.
        • Gensini G.F.
        • Casini A.
        Mediterranean diet and health status: an updated meta-analysis and a proposal for a literature-based adherence score.
        Public Health Nutr. 2014; 17: 2769-2782
        • United Nations Educational, Scientific and Cultural Organization
        Mediterranean Diet: Cyprus, Croatia, Spain, Greece, Italy, Morocco and Portugal Inscribed in 2013 (8.COM) on the Representative List of the Intangible Cultural Heritage of Humanity.
        UNESCO, Paris2013
        • Willett W.C.
        • Sacks F.
        • Trichopoulou A.
        • et al.
        Mediterranean diet pyramid: a cultural model for healthy eating.
        Am J Clin Nutr. 1995; 61: 1402S-1406S
        • León-Muñoz L.M.
        • Guallar-Castillón P.
        • Graciani A.
        • et al.
        Adherence to the Mediterranean diet pattern has declined in Spanish adults.
        J Nutr. 2012; 142: 1843-1850
        • Garcia-Meseguer M.J.
        • Burriel F.C.
        • Garcia C.V.
        • Serrano-Urrea R.
        Adherence to Mediterranean diet in a Spanish university population.
        Appetite. 2014; 78: 156-164
        • Barrington W.E.
        • White E.
        Mortality outcomes associated with intake of fast-food items and sugar-sweetened drinks among older adults in the Vitamins and Lifestyle (VITAL) study.
        Public Health Nutr. 2016; 19: 3319-3326
        • Pan A.
        • Sun Q.
        • Bernstein A.M.
        • et al.
        Red meat consumption and mortality: results from 2 prospective cohort studies.
        Arch Intern Med. 2012; 172: 555-563
        • Rohrmann S.
        • Overvad K.
        • Bueno-de-Mesquita H.B.
        • et al.
        Meat consumption and mortality--results from the European Prospective Investigation into Cancer and Nutrition.
        BMC Med. 2013; 11: 63
        • Schwingshackl L.
        • Schwedhelm C.
        • Hoffmann G.
        • et al.
        Food groups and risk of all-cause mortality: a systematic review and meta-analysis of prospective studies.
        Am J Clin Nutr. 2017; 105: 1462-1473
        • Guo J.
        • Astrup A.
        • Lovegrove J.A.
        • Gijsbers L.
        • Givens D.I.
        • Soedamah-Muthu S.S.
        Milk and dairy consumption and risk of cardiovascular diseases and all-cause mortality: dose-response meta-analysis of prospective cohort studies.
        Eur J Epidemiol. 2017; 32: 269-287
        • Moreira P.V.
        • Baraldi L.G.
        • Moubarac J.C.
        • et al.
        Comparing different policy scenarios to reduce the consumption of ultra-processed foods in UK: impact on cardiovascular disease mortality using a modelling approach.
        PLoS One. 2015; 10: e0118353
        • Moreira P.V.
        • Hyseni L.
        • Moubarac J.C.
        • et al.
        Effects of reducing processed culinary ingredients and ultra-processed foods in the Brazilian diet: a cardiovascular modelling study.
        Public Health Nutr. 2018; 21: 181-188
        • Fardet A.
        Minimally processed foods are more satiating and less hyperglycemic than ultra-processed foods: a preliminary study with 98 ready-to-eat foods.
        Food Funct. 2016; 7: 2338-2346
        • Nagai M.
        • Ohkubo T.
        • Miura K.
        • et al.
        Association of Total Energy Intake with 29-Year Mortality in the Japanese: NIPPON DATA80.
        J Atheroscler Thromb. 2016; 23: 339-354
        • Chien K.L.
        • Lin H.J.
        • Hsu H.C.
        • et al.
        Comparison of predictive performance of various fatty acids for the risk of cardiovascular disease events and all-cause deaths in a community-based cohort.
        Atherosclerosis. 2013; 230: 140-147
        • Kiage J.N.
        • Merrill P.D.
        • Robinson C.J.
        • et al.
        Intake of trans fat and all-cause mortality in the Reasons for Geographical and Racial Differences in Stroke (REGARDS) cohort.
        Am J Clin Nutr. 2013; 97: 1121-1128
        • Tasevska N.
        • Park Y.
        • Jiao L.
        • Hollenbeck A.
        • Subar A.F.
        • Potischman N.
        Sugars and risk of mortality in the NIH-AARP Diet and Health Study.
        Am J Clin Nutr. 2014; 99: 1077-1088
        • de Souza R.J.
        • Mente A.
        • Maroleanu A.
        • et al.
        Intake of saturated and trans unsaturated fatty acids and risk of all cause mortality, cardiovascular disease, and type 2 diabetes: systematic review and meta-analysis of observational studies.
        BMJ. 2015; 351: h3978
        • Wakai K.
        • Naito M.
        • Date C.
        • Iso H.
        • Tamakoshi A.
        • JACC Study Group
        Dietary intakes of fat and total mortality among Japanese populations with a low fat intake: the Japan Collaborative Cohort (JACC) Study.
        Nutr Metab (Lond). 2014; 11: 12
        • Guasch-Ferré M.
        • Babio N.
        • Martínez-González M.A.
        • et al.
        Dietary fat intake and risk of cardiovascular disease and all-cause mortality in a population at high risk of cardiovascular disease.
        Am J Clin Nutr. 2015; 102: 1563-1573
        • Graudal N.
        • Jürgens G.
        • Baslund B.
        • Alderman M.H.
        Compared with usual sodium intake, low- and excessive-sodium diets are associated with increased mortality: a meta-analysis.
        Am J Hypertens. 2014; 27: 1129-1137
        • Carocho M.
        • Barreiro M.F.
        • Morales P.
        • Ferreira I.C.F.R.
        Adding molecules to food, pros and cons: a review on synthetic and natural food additives.
        Compr Rev Food Sci Food Safety. 2014; 13: 377-399
        • Costa C.D.S.
        • Flores T.R.
        • Wendt A.
        • Neves R.G.
        • Assunção M.C.F.
        • Santos I.S.
        Sedentary behavior and consumption of ultra-processed foods by Brazilian adolescents: Brazilian National School Health Survey (PeNSE), 2015.
        Cad Saude Publica. 2018; 34: e00021017
        • Martínez-Gómez D.
        • Guallar-Castillón P.
        • León-Muñoz L.M.
        • López-García E.
        • Rodríguez-Artalejo F.
        Combined impact of traditional and non-traditional health behaviors on mortality: a national prospective cohort study in Spanish older adults.
        BMC Med. 2013; 11: 47
        • Gibney M.J.
        • Forde C.G.
        • Mullally D.
        • Gibney E.R.
        Ultra-processed foods in human health: a critical appraisal.
        Am J Clin Nutr. 2017; 106: 717-724

      Linked Article

      • Ultraprocessed Foods and Public Health: A Need for Education
        Mayo Clinic ProceedingsVol. 94Issue 11
        • Preview
          Ultraprocessed foods, which are made from food constituents but without any identifiable intact food in them, contain coloring chemicals, stabilizing substances, flavoring agents, and other additives that imitate or intensify the sensory qualities of foods or culinary preparations. They are also heavily loaded with free sugars, fats, salt, synthetic antioxidants, preservatives, and a variety of other chemical additives. Examples of ultraprocessed foods include sugar-sweetened beverages, sugared milk, fruit drinks, fast foods (eg, sausages, burgers), cookies, candy sweets, and savory packaged snacks.
        • Full-Text
        • PDF