André Krumel PortellaResearch & ideas

The thrifty behavior hypothesis

How early conditions may shape the way we seek, choose, and respond to food as well as the way our bodies use energy.

By André Krumel Portella, MD, PhD · Research Associate, Douglas Research Centre

From the body to behavior

The thrifty phenotype hypothesis proposed that conditions during development can shape how the body handles energy later in life. Our research asked whether those same early conditions may also shape behavior: how people respond to food rewards, what they prefer to eat, and how active they are. These behavioral changes could matter most when food is plentiful and heavily marketed.

This sits within developmental origins of health and disease (DOHaD): research on how early conditions relate to later health. Here, “thrifty” describes a proposed response to scarcity.

C. Nicholas Hales and David Barker’s 1992 paper is the starting point. It proposed that poor nutrition during development could leave lasting changes in how the body manages glucose and energy. Subsequent work by Barker, Hales, and colleagues linked reduced fetal growth to a higher risk of hypertension and type 2 diabetes later in life. My colleagues and I asked the next question: might development also alter the decisions and responses that determine how energy enters the body?

Our papers have used the terms ‘thrifty eating hypothesis’, ‘thrifty eating phenotype’, ‘thrifty behavior’, and ‘thrifty behavior phenotype’. The broader term includes activity and other behaviors as well as eating.

How the pieces may connect

  1. Early conditions

    Nutrition, growth, stress, and other conditions before birth and in early life

  2. Developing systems

    Possible changes in metabolism and in brain systems involved in reward, appetite, and self-regulation

  3. Later behavior

    Differences in food choices, response to tempting food, and activity may emerge

  4. Later environment

    Food availability, prices, advertising, family circumstances, and opportunities to be active influence what those tendencies mean for health

This sequence is a research model. It does not mean every step has been proven in people, or that one early exposure fixes a person’s future.

Evidence across research groups

Authorship and research lineage matter when judging evidence. The first set includes papers I coauthored. Later work from Patrícia Pelufo Silveira’s group did not include me as an author. Separate groups have reported related findings, as well as results that differ.

Research I coauthored

Observational studies and reviews that developed the behavioral extension.

  1. 2009

    A signal in adult food choices

    Severe intrauterine growth restriction is associated with higher spontaneous carbohydrate intake in young women.

    In a cohort of young women, severe restriction of growth before birth was associated with a higher spontaneous share of carbohydrates in the diet. This was an association, not proof that restricted growth caused the preference.

    Read the publication summary
  2. 2012

    The hypothesis was stated and tested

    Preliminary evidence for an impulsivity-based thrifty eating phenotype.

    A study of preschool children proposed the ‘thrifty eating hypothesis’. In girls, performance on a snack-delay task at age three was associated with fat intake and body mass index at age four. The study described its evidence as preliminary.

    Read the publication summary
  3. 2012

    A broader account of food preference

    Effects of in utero conditions on adult feeding preferences.

    A review led by Portella brought together human and animal studies of conditions in the womb and later food preferences. It proposed ‘thrifty behavior’ as a possible behavioral facet of the thrifty phenotype.

    Read the publication summary
  4. 2018

    Evidence in school-age children

    Low birth weight is associated with increased fat intake in school-aged boys.

    In a study of children aged six to twelve, lower birth weight was associated with more dietary fat and more energy-dense food among boys. The pattern was not the same in girls, and total calorie intake did not differ.

    Read the publication summary
  5. 2019

    Beyond a single eating trait

    Multi-behavioral obesogenic phenotypes among school-aged boys and girls along the birth weight continuum.

    A study led by Portella examined activity, screen time, eating, sleep, and body size across the birth-weight range. Associations varied by sex and behavior; only emotional eating in girls supported the tested pathway between birth weight and body size. The broader model remains a research hypothesis.

    Read the publication summary

Silveira-group follow-ups

New tests led by a central collaborator, without my authorship. These extend the program but are not independent replications.

  1. 2015

    An experimental test in rats

    Intrauterine growth restriction increases the preference for palatable foods and affects sensitivity to food rewards in male and female adult rats

    In a rat model, restricted growth before birth was followed by stronger preference for palatable food and differences in measures of the brain's reward system. This tests a possible mechanism in animals; it does not establish that the same mechanism explains human food choices.

    Read the original study
  2. 2016

    Food cues and eating in adolescents

    Impulsivity-based thrifty eating phenotype and the protective role of n-3 PUFAs intake in adolescents

    Among 48 adolescents and young adults, birth-size measures related to brain responses to food images, and omega-3 blood levels related to eating in response to outside food cues in the growth-restricted subgroup. The smaller brain-scan subset and observational design limit causal claims.

    Read the original study
  3. 2022

    A two-cohort food-choice test

    Thrifty-Eating Behavior Phenotype at the Food Court – Programming Goes Beyond Food Preferences

    Brazilian adolescents born smaller spent less to buy a snack and showed different brain connectivity. In the Canadian cohort, the food-choice difference was not statistically significant; the brain-connectivity findings also differed. This is a valuable, mixed follow-up rather than a clean replication.

    Read the original study

Evidence from other groups

Separate cohorts ask related questions. Their measures and populations differ, so these are convergent or contrasting evidence rather than exact repetitions of our studies.

  1. 2008

    Dutch famine cohort

    Prenatal exposure to the Dutch famine is associated with a preference for fatty foods and a more atherogenic lipid profile

    Among 730 adults, those exposed to famine early in pregnancy were more likely to fall in the highest-fat-diet group decades later. Average fat intake did not differ significantly across groups. This separate line of evidence predates our explicit eating hypothesis.

    Read the original study
  2. 2012

    Helsinki birth cohort

    Body size at birth is associated with food and nutrient intake in adulthood

    In 1,797 older adults, smaller size at birth was associated with lower later fruit and berry intake and a higher share of calories from fat. This is a separate cohort with a related pattern, not a direct repeat of the preschool snack-delay study.

    Read the original study
  3. 2021

    A contrasting U.S. preschool finding

    The association between birth weight and fat, sugar, and vegetable consumption in a national sample of U.S. preschool age children

    In 1,782 U.S. preschool children, the predicted adverse diet pattern was not observed. Boys with lower birth weight ate less solid fat and sugar and more vegetables. The finding shows that age, setting, and measures can change the picture.

    Read the original study

What the evidence can establish

  • The idea extends Hales and Barker’s account of metabolism to behavior. It does not replace the original hypothesis or suggest that metabolism alone is unimportant.
  • Silveira-group studies extend the research program without being independent replication. Other cohorts test related questions, often with different ages, exposures, and measures; their findings include support, differences, and non-replication.
  • Findings differ by age, sex, setting, and the behavior measured, suggesting complex interactions among life trajectory, genetic susceptibility or resilience, and the type and intensity of early exposure.
  • Food environments and social conditions matter. The hypothesis asks how early development may interact with those later surroundings, not whether a person is destined to eat in a particular way.

Where the research needs to go

  • Which early conditions matter most, and during which periods of development?
  • Which changes in brain and hormone systems actually connect early conditions to later choices?
  • How do food access, stress, family practices, and neighborhood environments change the outcome?
  • Can supportive environments or interventions alter these pathways over time?

Original sources

Hales and Barker’s original hypothesis and the publications on which this account is based. Each citation links to the publisher or publication record.

  1. Foundational hypothesisHales CN, Barker DJ. Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis. Diabetologia. 1992;35:595–601. Source record
  2. Foundational follow-upBarker DJ, Hales CN, Fall CH, Osmond C, Phipps K, Clark PM. Type 2 (non-insulin-dependent) diabetes mellitus, hypertension and hyperlipidaemia (syndrome X): relation to reduced fetal growth. Diabetologia. 1993;36:62–67. Source record
  3. Human observational studyBarbieri MA, Portella AK, Silveira PP, Bettiol H, Agranonik M, Silva AA, Goldani MZ. Severe intrauterine growth restriction is associated with higher spontaneous carbohydrate intake in young women. Pediatr Res. 2009;65(2):215-220. Source record Plain-language summary
  4. Human observational study and explicit hypothesisSilveira PP, Agranonik M, Faras H, Portella AK, Meaney MJ, Levitan RD; MAVAN Study Team. Preliminary evidence for an impulsivity-based thrifty eating phenotype. Pediatr Res. 2012;71:293–298. Source record Plain-language summary
  5. Review and ‘thrifty behavior’ framingPortella AK, Kajantie E, Hovi P, Desai M, Ross MG, Goldani MZ, Roseboom TJ, Silveira PP. Effects of in utero conditions on adult feeding preferences. J Dev Orig Health Dis. 2012;3(3):140-152. Source record Plain-language summary
  6. Review of possible brain and behavior pathwaysPortella AK, Silveira PP. Neurobehavioral determinants of nutritional security in fetal growth-restricted individuals. Ann N Y Acad Sci. 2014;1331:15-33. Source record Plain-language summary
  7. Human observational studyBischoff AR, Portella AK, Paquet C, Dalle Molle R, Faber A, Arora N, Levitan RD, Silveira PP, Dube L. Low birth weight is associated with increased fat intake in school-aged boys. Br J Nutr. 2018;119(11):1295-1302. Source record Plain-language summary
  8. Human observational study of multiple behaviorsPortella AK, Paquet C, Bischoff AR, Dalle Molle R, Faber A, Moore S, et al. Multi-behavioral obesogenic phenotypes among school-aged boys and girls along the birth weight continuum. PLoS One. 2019;14:e0212290. Source record Plain-language summary
  9. Silveira group · animal experimentDalle Molle R, Laureano DP, Alves MB, Reis TM, Desai M, Ross MG, Silveira PP. Intrauterine growth restriction increases the preference for palatable foods and affects sensitivity to food rewards in male and female adult rats. Brain Res. 2015;1618:41–49. Source record
  10. Silveira group · human observational studyReis RS, Dalle Molle R, Machado TD, et al. Impulsivity-based thrifty eating phenotype and the protective role of n-3 PUFAs intake in adolescents. Transl Psychiatry. 2016;6:e755. Source record
  11. Silveira group · two-cohort human studyDalle Molle R, de Mendonça Filho EJ, Minuzzi L, et al. Thrifty-Eating Behavior Phenotype at the Food Court—Programming Goes Beyond Food Preferences. Front Endocrinol. 2022;13:882532. Source record
  12. Separate Dutch cohortLussana F, Painter RC, Ocke MC, Buller HR, Bossuyt PM, Roseboom TJ. Prenatal exposure to the Dutch famine is associated with a preference for fatty foods and a more atherogenic lipid profile. Am J Clin Nutr. 2008;88:1648–1652. Source record
  13. Separate Finnish cohortPerälä MM, Männistö S, Kaartinen NE, et al. Body size at birth is associated with food and nutrient intake in adulthood. PLoS One. 2012;7:e46139. Source record
  14. Separate U.S. study · contrasting resultPotis K, Youngers EH, Tandy TK, Takemoto E, Boone-Heinonen J. The association between birth weight and fat, sugar, and vegetable consumption in a national sample of U.S. preschool age children. Matern Child Health J. 2021;25:1050–1056. Source record