Transgenerational effects of prenatal exposure to the 1944–45 Dutch famine


Correspondence: Dr M Veenendaal, Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Centre, PO Box 22660, 1100 DD Amsterdam, the Netherlands. Email



We previously showed that maternal under-nutrition during gestation is associated with increased metabolic and cardiovascular disease in the offspring. Also, we found increased neonatal adiposity among the grandchildren of women who had been undernourished during pregnancy. In the present study we investigated whether these transgenerational effects have led to altered body composition and poorer health in adulthood in the grandchildren.


Historical cohort study.


Web-based questionnaire.


The adult offspring (F2) of a cohort of men and women (F1) born around the time of the 1944–45 Dutch famine.


We approached the F2 adults through their parents. Participating F2 adults (n = 360, mean age 37 years) completed an online questionnaire.

Main outcome measures

Weight, body mass index (BMI), and health in F2 adults, according to F1 prenatal famine exposure.


Adult offspring (F2) of prenatally exposed F1 fathers had higher weights and BMIs than offspring of prenatally unexposed F1 fathers (+4.9 kg, = 0.03; +1.6 kg/m², = 0.006). No such effect was found for the F2 offspring of prenatally exposed F1 mothers. We observed no differences in adult health between the F2 generation groups.


Offspring of prenatally undernourished fathers, but not mothers, were heavier and more obese than offspring of fathers and mothers who had not been undernourished prenatally. We found no evidence of transgenerational effects of grandmaternal under-nutrition during gestation on the health of this relatively young group, but the increased adiposity in the offspring of prenatally undernourished fathers may lead to increased chronic disease rates in the future.


Human and animal studies have shown that the fetal environment can affect the health of an individual throughout their lifetime. This is thought to reflect programming, the process by which the same genotype can give rise to several phenotypes depending on the early environment. We have previously reported that individuals exposed to the Dutch famine of 1944–45 in utero have increased rates of cardiovascular disease, type-2 diabetes, and breast cancer.[1-3] There are indications that these programming effects may be transmitted across generations, as has been shown in a number of rodent studies. Feeding rats a protein-restricted diet during gestation not only resulted in higher blood pressure and endothelial dysfunction in the offspring, but also in the grand-offspring.[4] A low-protein diet during pregnancy led to insulin resistance in the adult male and female F2 offspring.[5, 6] There is evidence suggesting that glucose metabolism in the F3 generation is also affected by F0 under-nutrition.[7] In mice, maternal general under-nutrition during gestation led to reduced birthweight, impaired glucose tolerance, and obesity in the F1 and F2 generations in a gender-specific manner.[8] Transgenerational effects are also seen in humans: for instance, the offspring of people born with low birthweight have an increased risk of having babies with reduced birthweight themselves;[9] moreover, maternal diet during the F0 pregnancy seems to affect the F2 birthweight, independently of the F1 birthweight.[10, 11]

Studies reporting on transgenerational effects of prenatal under-nutrition in humans are scarce. A historical study of three generations in Overkalix, Sweden, reported that a limited food supply for the grandparents influenced their grandchildren's later mortality and disease risk in a sex-specific manner, in part operating exclusively through the paternal line.[12]

We have previously demonstrated that prenatal famine exposure produced transgenerational effects. We found that F1 women who had experienced famine as fetuses had F2 babies with increased neonatal adiposity and poorer adult health.[13] This study, however, was based on parents' recall of their offspring's size at birth and later health, which may have led to a level of inaccuracy. In the study reported here we contacted the offspring directly to measure their body composition and investigate their health.


Participants and selection

The Dutch Famine Birth Cohort consists of 2414 men and women born between 1 November 1943 and 28 February 1947 as term singletons in the Wilhelmina Gasthuis, a local hospital in Amsterdam. The selection procedures and loss to follow-up until 2002 have been described in detail elsewhere.[14, 15] Cohort members were eligible for participation in this study if they lived in the Netherlands on 1 September 2008, and if their address was also known to us. From 2002 onwards, 31 people had died, six had emigrated, 11 had an unknown address, and eight had requested their address to be removed from our database. In total, 1372 eligible individuals were invited to participate in the current study (Figure 1).[16] The study was approved by the local Medical Ethics Committee and carried out in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Figure 1.

Flow diagram of the experimental design.

Exposure to famine

We defined famine exposure according to the official daily food rations for adults.[17] A person was considered to be prenatally exposed to famine if the average daily food ration of the mother during any 13-week period of gestation contained less than 1000 calories. Therefore, babies born between 7 January 1945 and 8 December 1945 had been exposed in utero. People born before 7 January 1945 and conceived and born after 8 December 1945 were considered as unexposed to famine in utero, and acted as control groups (Figure 2).

Figure 2.

The Dutch famine birth cohort: timing of the Dutch famine (top bar with dates), as related to the timing of famine exposure according to the timing of birth of the cohort members (bottom bar).


We studied two generations. The F1 generation included the men and women from the Dutch famine birth cohort, born between November 1943 and February 1947. The F2 generation (of grandchildren) included all of the offspring of F1 men or women.

Data collection

Information about the mother, the course of the pregnancy, and the size of the baby at birth (F1) was extracted from medical birth records.[15] Previously, at age 58 years, F1 participants visited the clinic or were seen at home, where F1 weight was measured with Seca scales or Tefal portable scales, and height was measured using a fixed or portable stadiometer. We asked them about the birthweight, birth length, and gestational age at delivery of their children (F2).[13] We calculated the ponderal index (birth weight/birth length³) as a measure of neonatal adiposity.

In the current study, all F2 participants (grandchildren) were asked to give information concerning their medical history, lifestyle, and children. Data were collected by means of a standardised questionnaire that was filled out at home by the participants, either on paper or using a web-based form. The questionnaire included questions on height, weight, smoking, alcohol consumption, and exercise behaviour. We obtained information about symptoms or a history of cardiovascular, pulmonary, psychiatric, and metabolic disease, and medication use. The questions concerning cardiovascular disease included the Rose questionnaire.[18] Questions were combined to achieve categories relating to cardiovascular disease, pulmonary disease, hay fever, eczema, cholesterol, diabetes, and hypertension. For each condition, questions were phrased as ‘has a doctor ever diagnosed (condition)’ or ‘has a doctor ever prescribed medication for (condition)’.

Statistical methods

We used linear regression for continuous variables and logistic regression for dichotomous variables to compare (grand-) maternal, birth, and adult outcomes of those exposed and those unexposed to famine during gestation, and also to compare the offspring of these groups. To take into account the correlation of characteristics between siblings, we used mixed models to analyse the association between F1 famine exposure during different stages of gestation and F2 birth and health characteristics. F1 body mass index (BMI), F1 weight, and F2 birthweight were missing in a relatively large number of participants (16, 16, and 28%, respectively). To deal with these missing values, we applied multiple imputation using the linear regression method. We adjusted for possible confounding factors, such as F2 (grandchildren) age, sex, and birthweight, and F1 BMI and weight. We report the F2 (grandchildren) characteristics stratified according to F1 sex. We used spss 19.0 (SPSS Inc., Chicago, IL, USA) for all analyses.


In the eligible population of 1372 cohort members (F1), 483 F2s (grandchildren) were willing to participate in the current study. Of these, 360 (74.5%) completed the questionnaire (Figure 1). The mean age at participation was 37 years (range 18–47 years). In total, 135 males and 225 females participated. Birthweight or gestational age did not differ between F1 participants and non-participants (> 0.8).

The F2 offspring of F1 exposed men were 2.1 years younger than offspring from unexposed F1 men (Table 1). The birth characteristics did not differ between F2 offspring of the exposed and unexposed F1 men. The F2 offspring of F1 exposed women had a higher ponderal index at birth, compared with the F2 offspring of unexposed F1 women (Table 1). Exposed F1 men and women were comparable with unexposed F1 men and women with regards to anthropometric measures at age 58 years.

Table 1. Characteristics of F2 participants according to F1 gender
  F1 male F1 female
ExposedUnexposedAll ± SDExposedUnexposedAll ± SD
  1. a

    < 0.05.

  2. b

    < 0.05 after correction for age and sex of F2.

  3. c

    Geometric mean and SD.

n 5299151106103209
Sex (% male)4235a384628a37
Age (years)33.936.0a35.3 ± 4.736.738.3a37.5 ± 4.6
Birthweight (g)335133423345 ± 826337432453313 ± 707
Birth length (cm)50.549.449.8 ± 6.450.250.550.3 ± 3.4
Ponderal index (kg/m³)27.626.326.7 ± 6.226.7a24.725.8 ± 4.8
Adult length (m)1.761.741.75 ± 0.11.761.731.75 ± 0.1
Adult weight (kg)78.8b73.575.3 ± 14.479.178.979.0 ± 18.0
BMI (kg/m²)c25.2b23.824.3 ± 3.825.025.725.3 ± 5.3

F2 adult body composition


Offspring of exposed F1 fathers had a higher BMI than offspring of unexposed F1 fathers (+1.6 kg/m², 95% CI 0.5–2.6), after adjusting for F2 (grandchildren) sex and age (P = 0.006). Adding F2 (grandchildren) birthweight and F1 BMI to the model did not change the association (+1.5 kg/m², 95% CI 0.4–2.5, P = 0.005). There was no effect of maternal F1 exposure to famine on the BMI of the offspring (unadjusted −0.69 kg/m², 95% CI −3.5 to 2.1, P = 0.36; adjusted −0.60 kg/m², 95% CI −1.7 to 0.5, P = 0.30).


Offspring of exposed F1 fathers were heavier than offspring of unexposed F1 fathers (+4.9 kg, 95% CI 0.8–9.1, P = 0.03), after adjustment for F2 sex and age. This effect remained when adjustments were made for F2 birthweight and F1 weight (+4.5 kg, 95% CI 0.9–8.1, P = 0.01). Offspring of F1 famine-exposed mothers were not heavier than F2 offspring of unexposed mothers (unadjusted −1.7 kg, 95% CI −5.5 to 2.0, P = 0.36; adjusted −0.7 kg, 95% CI −4.5 to 3.1, P = 0.72).

F2 self-reported health

We did not find differences in the prevalence of cardiovascular and pulmonary disease, hay fever, eczema, elevated cholesterol, diabetes or hypertension between offspring of men and women exposed to famine during gestation, and offspring of unexposed men and women (all > 0.1; Table 2).

Table 2. Prevalence of F2 self-reported diseasea according to F1 gender
  F1 exposed F1 unexposed All
  1. a

    Defined as answering in the affirmative to questions phrased as ‘has a doctor ever diagnosed’ or ‘has a doctor ever prescribed medication for’ the different conditions.

F2 of F1 men
n 5299151
Hay fever%19.630.626.8
F2 of F1 women
n 106103209
Hay fever%20.618.019.3


In this study we found that offspring of fathers that had been exposed to famine prenatally were heavier and had a higher BMI than offspring of unexposed fathers. This effect remained after adjustment for birthweight, and for paternal weight and BMI. We could not demonstrate a transgenerational effect on health among the offspring of prenatally exposed men or women.

These findings fit within the growing body of evidence that transgenerational non-genomical inheritance specifically takes place in the paternal line. For instance, embryonic exposure to the endocrine disruptor vinclozolin increased a variety of adult-onset diseases in the subsequent generations specifically through the paternal line.[19, 20] In a study exploring the mechanism underlying the transgenerational effects of vinclozolin exposure, it appeared to promote epigenetic reprogramming of the male germ line, correlating with transgenerational alterations in the testis transcriptome in subsequent generations.[21] Other studies have supported this: vinclozolin exposure of pregnant mice altered the imprinting status of genes in the sperm of the offspring.[22] Dietary exposures have also been shown to produce transgenerational effects through the male line: in Sprague–Dawley rats, fathers who consumed a diet high in fat diet had female offspring with impaired glucose tolerance and insulin secretion.[23] In mice, a paternal low-protein diet induced altered expression of genes involved in lipid and cholesterol metabolism in the liver of the offspring, compared with offspring of control-fed male mice. These expression differences were thought to be the result of alterations of the epigenome.[24]

In this cohort, we reported no epigenetic changes after prenatal exposure to famine,[25] but others have found evidence that prenatal exposure to the Dutch famine induced epigenetic alterations in the F1 generation that persist throughout life.[26, 27] It is unknown whether these epigenetic alterations are transmitted to the next generation, and whether they are sex-specific, so that they could explain the transgenerational effects that we found only in the paternal line. Although epigenetics is a very likely mechanism explaining transgenerational effects in model organisms, in humans other mechanisms may also play a role.

The transgenerational effects we found may also have been transmitted through environmental factors such as food preferences and physical activity. Although the transgenerational effects reported here were also seen after adjustment for paternal BMI and weight, parental obesity is known to be associated with the level of overweight and obesity in children,[28] and it significantly alters the risk of obesity in adulthood.[29] Parents are responsible for the quality and availability of the food in the home, and their food habits will be adopted by their children.[30] Children from obese or overweight families have been shown to have a higher preference for fatty foods, a lower liking of vegetables, and lower physical activity than children from lean families.[31] We previously reported that people who were conceived during the Dutch famine were twice as likely to consume a high-fat diet and to have a tendency to be less physically active.[32] Transgenerational propagation of unhealthy lifestyle patterns may have contributed to the increased weight and BMI of the F2 offspring of exposed F1 men.

Previously, we found that participants who were themselves exposed to famine prenatally rated their offspring's health more often as poor than did unexposed participants.[13] We set out in this study to investigate the self-reported health of the F2 (grandchildren) offspring, but we could not demonstrate any effects of F1 famine exposure in utero on the health of the F2 generation. A number of methodological issues may explain the fact that we did not find an effect on the health of the F2 offspring (grandchildren). The group studied consisted of only 360 people, which limits the power to detect an effect. The birth characteristics of the F1 participants did not differ from the non-participants: therefore, we have no evidence of selection bias occurring at the level of F1 participants informing their children about the study. Also, we do not feel that selection played a role on the level of the F2 offspring (grandchildren), as it would be unlikely that only healthy or unhealthy F2 (grandchildren) participated in this study. Still, we are limited by the design of this study, in which we approached the cohort members (F1) to inform their children about the study and ask them to participate. Although a reminder was sent after the initial study invitation, we do not known whether all eligible F2 participants (grandchildren) were informed about the study, because ethics regulations forced us to contact the F2 participants (grandchildren) through their parents.

The statistical power to detect effects was further reduced because the mean age of the F2 offspring (grandchildren) was 37 years, which is relatively young when studying the prevalence of chronic disease. Also, our analyses are based on self-reported health questionnaires, which are a fairly crude measure of health.

In conclusion, we did not find a transgenerational effect of prenatal famine exposure on the health of grandchildren in this study, but we did find increased weight and BMI among F2 offspring (grandchildren) of men exposed to famine in utero. These results warrant further follow-up of the health of the F2 generation as they age, as their increased adiposity may predispose them to increase disease risk later on. Also, they suggest that transmission through the paternal line may occur, and so future studies in this cohort will investigate mechanisms that may elucidate the biological processes that underlie these findings.

Disclosure of interests

None of the authors report any conflicts of interest.

Contribution to authorship

RCP and TJR set up the study. MVEV was responsible for data collection, performed the analyses, and drafted the article. All authors contributed to the interpretation of the results and took part in revising the article.

Details of ethical approval

The study was approved by the Medical Ethics Committee at the Academic Medical Centre, Amsterdam, the Netherlands (Mec 08/26). Written informed consent was obtained from all participants in the study.


This work was funded by the Netherlands Heart Foundation (grant number 2007B083).


We gratefully thank the participants for participating in this study.

Reviewer's commentary on ‘Transgenerational effects of prenatal exposure to the 1944–45 Dutch famine’

This study looked at the grandchildren of adults who were both maternally and paternally exposed to the Dutch famine. The results are seemingly in contrast to earlier findings from this veteran research team on maternal transmission of risk for adiposity and poor health to the F2 generation (Painter RC et al., BJOG 2008;115:1243–49). In the current study, an effect on adiposity was seen but only in paternally exposed individuals and effects on combined aspects of poor health later in life were not demonstrated. Several factors may account for this discrepancy. As mentioned by the authors, the small sample size and the relatively young age of the F2 participants could be responsible. Additionally, the failure to see an increased risk for cardiovascular and metabolic disease may be related to a possible mismatch between poor nutrition during development and diet in later life (Godfrey KM et al., Pediatr Res 2007;61:5R–10R). This hypothesis, of developmental origins of adult disease, suggests that the environment during periods of developmental plasticity induces alterations in metabolic processes through epigenetic (non-chromosomal) modification of the transcriptome (RNA production), which can be transmitted from generation to generation. Nutrition is thought to be the major environmental factor responsible for epigenetic changes to the fetal transcriptome. Developmental plasticity provides a method to alter gene expression to produce a phenotype that is matched to the environment, thus conferring an adaptive advantage against starvation in the short term. When the resulting phenotype remains matched to the original inducing environment in later life, no increase in disease risk would be expected (Godfrey KM et al., Pediatr Res 2007;61:5R–10R). However, when the individual is exposed to malnutrition in utero and, conversely, later in life is exposed to our current obesogenic environment, characterised by ‘supersizing’ and the consumption of an abundance of high-fat, energy-dense, foods, combined with low physical activity, an increased risk for metabolic and cardiovascular disease may result. A growing body of evidence from animals and humans suggests that dietary exposures in utero can have effects on subsequent generations (McKay JA and Mathers JC, Acta Physiol 2011;202:103–18). Transgenerational effects are potentially transmitted through multiple epigenetic mechanisms, which may increase the risk of chronic disease. Although the current study did not evaluate epigenetic mechanisms, it is well accepted that epigenetic changes during development leave permanent marks on the transcriptome that can be passed from generation to generation without changing the DNA sequence. Epigenetic effects include primarily three different interacting mechanisms: (1) DNA methylation; (2) histone modification; and (3) non-coding microRNAs, which are responsible for gene expression during development and throughout life. DNA methylation is involved in several key processes, and alterations induced by DNA methylation enzymes are known to be affected by nutrition. This includes the availability of methyl donors and alteration of enzyme activity in response to folate, vitamin B6, vitamin B12, and choline (McKay JA and Mathers JC, Acta Physiol 2011;202:103–18). Negative effects of low-protein diets, mediated by the limited availability of methyl donors, have been shown to be mitigated by additional folate (Lillycrop KA et al., J Nutr 2005;135:1382–86). Histone modification, resulting in permanent epigenetic changes (known as marks) via post-translational methylation of lysine and arginine residues, and acetylation and ubiquination of serine residues, alters the chromatin structure and impacts gene translation. These changes work in concert with changes in microRNAs, which control about 30% of protein-encoding genes. Dietary factors including protein, ethanol, and vitamin E intakes have been shown to alter microRNA expression in humans and animals. Individuals exposed to the Dutch Hunger Winter (1944–45) in utero have been shown to have lower methylation of the insulin-like growth factor 2 (IGF2) gene, compared with non-exposed individuals (Heijmans BT et al., PNAS 2008;105:17046–49). Conversely, children whose mothers used periconceptional folate supplements had higher methylation of their IGF2 gene. Data from the Dutch famine, a ‘natural’ experiment that can never be duplicated, are important for providing insights into the effect of malnutrition during development, and the subsequent effects of diet on risk for disease.

Disclosure of interests

I have no conflicts of interest to disclose.

M Harris

Professor of Nutrition, Colorado State University, Fort Collins, CO 80523, USA