During pregnancy, a healthy, balanced diet, characterised by regular consumption of, e.g. vegetables, fruits, whole grains and fatty fish, is required for the health of women and their children in utero and later life(Reference Koletzko, Bauer and Bung1). A healthy diet contributes to fetal development and growth, prevents exceptional maternal weight gain during pregnancy and decreases the risk for gestational diabetes(Reference Mijatovic-Vukas, Capling and Cheng2), preterm delivery(Reference Englund-Ögge, Brantsæter and Sengpiel3,Reference Chia, Chen and Lai4) and increased birth weight(Reference Grandy, Snowden and Boone-Heinonen5). In newborns, an unbalanced diet in pregnancy impairs insulin, glucose and lipid metabolism and consequently increases the risk for CVD and type 2 diabetes in later life(Reference Holness, Langdown and Sugden6,Reference Koletzko, Brands and Grote7) . A higher intake of processed food during pregnancy has been shown to be positively associated with obesity during childhood(Reference Rohatgi, Tinius and Cade8). In contrast, inadequate micronutrient intake during pregnancy increased the risk of low birth weight and pre-eclampsia(Reference Keats, Haider and Tam9,Reference Hofmeyr, Lawrie and Atallah10) .
The diet composition of pregnant women is further related to how they eat (‘eating culture’, e.g. number of meals per day). Although the diet during pregnancy has been widely examined, information on the eating culture of pregnant women is scarce. It has been demonstrated that skipping meals during pregnancy was associated with lower circulating levels of EPA, DHA and β-carotene(Reference Shiraishi, Haruna and Matsuzaki11) and with reduced diet quality in low-income women(Reference Fowles, Stang and Bryant12).
It has been shown in further studies that diet is related to sociodemographic parameters such as age, gender and socio-economic background in children, adolescents(Reference Dubois, Farmer and Girard13,Reference Lien, Jacobs and Klepp14) , adults(Reference Mayén, Marques-Vidal and Paccaud15) and during pregnancy(Reference Northstone, Emmett and Rogers16–Reference Jardi, Aparicio and Bedmar19). More precisely, healthier eating behaviour during pregnancy is positively associated with a higher socio-economic status (SES)(Reference Wesolowska, Jankowska and Trafalska18), level of education(Reference Arkkola, Uusitalo and Kronberg-Kippilä17,Reference Wesolowska, Jankowska and Trafalska18) and increasing age(Reference Arkkola, Uusitalo and Kronberg-Kippilä17,Reference Wesolowska, Jankowska and Trafalska18) of the woman and negatively with unemployment(Reference Northstone, Emmett and Rogers16).
This project aimed to establish and validate a short questionnaire on diet composition and eating culture. The Composition and Culture of Eating ‘CoCu’ has already been validated for children and adolescents(Reference Poulain, Spielau and Vogel20). Here, we assessed the validity of the version for pregnant women. For this purpose, we compared the answers provided in CoCu with answers provided in a FFQ completed at the same time point. Predictive validity was not explored because the main aim of the questionnaire is to assess diet and not to predict further development. We further investigated whether given responses in 24th and 36th week of pregnancy (wp) are comparable. These results will provide interesting indications of changes in eating habits during pregnancy.
This questionnaire is intended to assess both diet composition and culture of eating in the areas of research and everyday clinical practice. Further, we suggest that it is necessary to understand how social-demographic factors (e.g. age and SES) of women of childbearing age might influence their food intake.
The knowledge of influencing factors can be used in further prevention programmes as women are particularly changing their lives during this time of life.
Materials and methods
Study participants and design
Data were collected between 2016 and 2020 in the framework of the LIFE Child study. The LIFE Child study is a longitudinal study investigating normal child development and the development of civilisation diseases from pregnancy to adulthood (NCT02550236)(Reference Poulain, Baber and Vogel21,Reference Quante, Hesse and Döhnert22) .
Within the study, pregnant women also pass a detailed study programme. They are invited to participate at two time points during pregnancy, once in the 24th wp and once in the 36th wp. Any queries were answered on the spot by study assistants.
CoCu was completed by 430 women in the 24th wp. Most analyses presented in this paper are based on this sample. Of these women, ninety-seven have completed the questionnaire at both time points during pregnancy. The analyses on agreement between responses were based on this subsample.
Study participants provided information on their education, their occupational position and their family net income for the assessment of SES. Based on this information, a SES score was derived(Reference Lampert, Müters and Stolzenberg23,Reference Winkler and Stolzenberg24) which ranged between 3 and 21, with a higher score indicating higher SES. Scores between 3 and 8·4 indicate a low SES, scores between 8·5 and 15·4 a middle SES and scores between 15·5 and 21 a high SES. In a representative sample, 60 % of participants would be expected to have a middle SES, 20 % a lower SES and another 20 % a higher SES(Reference Lampert, Müters and Stolzenberg23).
Measures
Composition and Culture of Eating pregnancy: composition and culture of eating
CoCu is a screening instrument developed to assess the composition of diets and the culture of eating in children and adults. For children, there exists two different versions of the questionnaire (self- and parent-reported version) which have already been described previously(Reference Poulain, Spielau and Vogel20). The CoCu pregnancy is nearly identical to the version for children but includes a further question about raw product consumption. It consists of two parts, a diet composition part and a culture of eating part. The composition of diet part contains fourteen questions on what women eat. In more detail, women indicate how many portions of different food products they eat per day (for fruits/vegetables, unsweetened milk products, sweetened milk products, sweetened beverages, whole-grain bread, white bread) or per week (for meat, fish, ready-made meals, fried potatoes, potatoes, rice/noodles, cakes, sweet or savoury snacks). Answers are given on a six-point scale ranging from 1 (‘never’) to 6 (‘>7 portions’). Reference portions are described in the text (e.g. ‘a fistful’) or illustrated using photographs.
The selection of food items was largely based on a FFQ that has been applied in the LIFE Child study for several years(Reference Stiegler, Sausenthaler and Buyken25). We summarised the inquired foods (e.g. apple, berry fruits, salad, carrots) from FFQ as categories (e.g. fruit/vegetables). The categories were included into the CoCu. The questions in the short questionnaire only cover the food category, not a specific food itself. Some of the food groups presented in the FFQ were not included in CoCu, as they were judged to be consumed too infrequently (e.g. oil, nuts), to reveal little about the healthiness of the subject’s diet (e.g. tea, water) or to be inappropriate for pregnant women (e.g. alcohol).
Importantly, the answers for the composition of diet part of CoCu pregnancy can be analysed in two phases. First, the consumption of each food product can be described in terms of portions/week. Second, the level of consumption of each food product can be categorised as either healthy (‘green’ or ‘10 points’), moderately healthy (‘yellow’ or ‘0 points’) or unhealthy (‘red’ or ‘–10 points’). The categorisation into ‘green’, ‘yellow’ and ‘red’ was adapted from CoCu for children and adolescents(Reference Poulain, Spielau and Vogel20), with specific parameters set for each food product (see online supplementary material, Supplemental 1). For example, for fruits/vegetables, answers of ‘4–5’ portions/week or more were categorised as ‘green’, answers of ‘2–3 portions/week’ were categorised as ‘yellow’ and answers of ‘never’ or ‘max.1/week’ were categorised as ‘red’. Three of the original food items (sweetened milk products, cakes, and sweet or savoury snacks) were aggregated into one item called ‘treats’.
As for the CoCu in children and adolescents(Reference Poulain, Spielau and Vogel20), a ‘Nutritional Health Score’ (NHS) is derived by summing the categorisations/points of all food items. This score may range from –120 (in cases where all responses were categorised as unhealthy (–10 points)) to +120 (in cases where all responses were categorised as healthy (+10 points)), with higher scores indicating healthier diets.
The second part of CoCu pregnancy (eating culture) comprises six questions on the value of meals. The women are asked how many and which meals they eat each day (first breakfast, second breakfast, lunch, afternoon snack, dinner), whether they use media while eating, or whether they eat unhealthy snacks between meals.
As raw products should be avoided during pregnancy(Reference Koletzko, Bauer and Bung26), the CoCu pregnancy questionnaire also comprises questions on the consumption of raw milk products and raw meat. The questions could be answered by either Yes or No. Furthermore, we asked if the women follow a certain diet (vegetarian, vegan, gluten-free, lactose-free or other). The entire questionnaire, in German (original) or English (translation), is available on request.
FFQ
The FFQ applied in the LIFE Child study provides a detailed overview of a woman’s diet during pregnancy(Reference Stiegler, Sausenthaler and Buyken25). Participants indicate not only the frequency of consumption of different food products (choosing from nine possible frequency levels ranging from ‘never’ to ‘more frequently than 4 times per day’) but also the portion size they usually consume (three to five different answer categories). The eighteen food groups (forty-two subgroups) are reflected in eighty-two different questions. In the LIFE Child study, the FFQ is completed in the 24th and 36th wp(Reference Poulain, Baber and Vogel21).
Statistical analysis
The concurrent validity of CoCu pregnancy was investigated by comparing responses in CoCu pregnancy with responses in the FFQ using Spearman correlations. All items were expressed as portions/week or weight/week (where it was not possible to identify a portions/week value). Each item of CoCu pregnancy was associated with the appropriate item(s) of the FFQ. If an item of CoCu pregnancy was represented by more than one item in the FFQ, the relative items of the FFQ were aggregated. Spearman rank correlations were calculated.
Retest reliability is usually checked by comparing answers from two study visits. This procedure is fraught with some problems during pregnancy, since eating habits may change (stronger than outside pregnancy) and may also be affected by other factors such as nausea and abdominal fullness. Here, we calculated two types of percentage agreement for answers given in the 24th (T1) and 36th wp (T2). In the first evaluation, agreement was defined as having chosen exactly the same response category at both time points. In the second analysis, having chosen adjacent response categories (e.g. ‘max. 1 portion/week’ in the 24th wp, and either ‘never’, ‘max. 1/week’ or ‘2–3 portions/week’ in the 36th wp) was also considered as agreement.
Associations between the healthiness of women’s diet (as reflected by the NHS), the consumption of raw products and eating culture (reflected by the different items of the culture of eating part), on the one hand, and age and SES, on the other, were assessed using multiple linear or logistic regression analyses. Possible relationships between the healthiness of diet (as dependent variable) and eating culture (as independent variable) were investigated using multiple linear regression analysis. Age and SES were included as control variables. In linear regression analyses, effect sizes were described in terms of standardised regression coefficients (β). In the logistic regression analyses, they were described in terms of OR.
Results
The present sample
The two samples included in the analyses (430 women participating in the 24th wp, ninety-seven women participating in the 24th and 36th wp) are described in Table 1. The mean age of women completing the questionnaire in the 24th wp was 31·3, ranging from 21 to 39. The SES was rather high, with only 1 % having a low SES but 52 % having a high SES. The education was high as well, with approximately half of the participating women having a University degree. The majority of women (>80 %) came from an urban area. Information on the course of pregnancy and previous pregnancies was available for 379 participants. Of these women, 111 reported at least one complication during the current pregnancy (e.g. bleeding, placental insufficiency, cervical weakness, anaemia, gestational diabetes), 107 reported one previous pregnancy and eighty-one reported two previous pregnancies. Fifty-six women reported at least one previous miscarriage, and nineteen reported at least one previous abortion.
Description of the study samples analysed in the present analyses
* Troubles might be a general disease, long-term medication, abusus, mental stress, bleeding, placenta previa, multiple births, hydramnion, oligohydramnios, uncertainties regarding date of birth, placental insufficiency, cervical weakness, premature labour, anaemia, urinary tract infection, indirect Coombs test positive, abnormal serum findings, protein excretion 1 %, hypertension, hypotension, oedema, gestational diabetes, adjustment anomaly or other troubles.
A comparison of women participating in the 24th and 36th wp (n 97) and those lost to follow-up (n 333) revealed no significant differences regarding age, the distribution of SES and education, the frequency of one or two previous pregnancies or the frequencies of miscarriages or abortions in one of the previous pregnancies (all P > 0·05). However, compared with women participating in the 24th and 36th wp, those lost to follow-up consumed significantly more treats (9·82 v. 8·31 portions per week, P = 0·015) and ready-made meals (0·64 v. 0·44 portions per week, P = 0·019). Regarding the other food items and the NHS, the two samples did not differ significantly (all P > 0·05).
Diet composition, raw products, specific diets and eating culture in the present sample
Table 2 shows the consumption of different food products in terms of mean portions per week. The food products consumed most frequently were fruits and vegetables (mean portions per week = 21·05). The food products consumed least frequently were fried potatoes, ready-made meals and fish (all fewer than once per week, see Table 2). For all food products, the mean portions per week were categorised as either healthy (green) or moderately healthy (yellow). Only for sweetened beverages and treats, the mean responses of 8·28 and 9·48 portions per week indicated a tendency to moderately healthy consumptions (indicated by a daily consumption of more than one portion per day).
Diet composition, raw products, specific diets and eating culture in the present sample (24th wp) (n 430)
* Treats summarise sweetened milk products, cakes and snacks.
Nearly 16 % of the participating pregnant women reported eating some raw products, where raw meat was consumed more frequently than raw milk products (see Table 2).
Nearly 13 % of pregnant women reported that they eat according to a certain diet. Vegetarian and lactose-free diets were more frequent than vegan and gluten-free diets (see Table 2).
The average responses to the items of the eating culture part of the questionnaire are also displayed in Table 2. Pregnant women reported having on average between 3 and 4 meals per day. Nearly 50 % used media while eating, and more than 50 % reported to snack between meals.
Concurrent validity
The concurrent validity of CoCu pregnancy was tested by comparing answers given in CoCu pregnancy with answers given in a more detailed FFQ. The results of these analyses are presented in Table 3. For each food product, the correlations between answers given in both questionnaires were positive and significant (all P < 0·001), although some showed weak correlations (white bread and sweetened milk products (rho = 0·32)). The strongest correlation was observed for the consumption of fish (rho = 0·61).
Concurrent validity: Spearman rank correlations between the answers given in Composition and Culture of Eating (CoCu) pregnancy and those given in the FFQ
* P < 0·001.
† Due to missing values in the items of the FFQ, different numbers of participants were included in the single analyses.
Agreement between responses given in the 24th and 36th week of pregnancy
Percentage agreement of responses was assessed in a subsample of pregnant women (n 97) who had completed the CoCu questionnaire at two time points, namely in the 24th and 36th wp. As can be seen in Table 4, between 35 (sweet/savoury snacks) and 71 % (fish, potatoes) of women had chosen the same response category at both time points. Furthermore, the results demonstrated that 90 (beverage sweetened) to 99 % (rice/noodles, fried potatoes, fish) of pregnant women selected the same or adjacent categories at both study visits.
Percentage of agreement between responses given in 24th (T1) and 36th (T2) wp
Associations between diet composition/raw products/eating culture and age/socio-economic status
Diet composition
The analysis revealed significant positive associations between the NHS and age (β = 0·11 (95 % CI 0·01, 0·20), P = 0·027) as well as SES (β = 0·21 (95 % CI 0·11, 0·30), P < 0·001). For pregnant women aged 22 years, the NHS estimated by the statistical model was 29. For pregnant women aged 40 years, in contrast, the model estimated a NHS of 56. The estimated average NHS for pregnant women of middle (score of 12) and high SES (score of 18) were 43 and 54, respectively.
Raw products
Pregnant women with a lower SES were more likely to eat raw products than women with a higher SES (OR = 0·89, P = 0·011). In the middle social stratum, the percentage of women eating some raw products was estimated 21 %, compared with 12 % in the high social stratum.
Eating culture
The analyses on associations between the different eating culture items and age and SES are shown in Table 5. They revealed small but significant positive association between the number of meals per day and the age (β = 0·12, P = 0·011) and the SES (β = 0·11, P = 0·025) of the participating women. Furthermore, older women and women with a higher SES were less likely to use media while eating than younger women and women with a lower SES (OR = 0·90 and 0·84, respectively, both P < 0·001). Finally, pregnant women with a lower SES were more likely to eat unhealthy snacks between meals than pregnant women with a higher SES (OR = 0·93, P = 0·027).
Associations (indicated by standardised regression coefficients or OR) between diet composition, raw products, eating culture and age and SES
All associations are adjusted for the other independent variable.
* P < 0·05,
** P < 0·01,
*** P < 0·001.
† Linear regression.
‡ Logistic regression, reference = no raw products/no media while eating/no unhealthy snacks between meals.
Relations between Nutritional Health Score and eating culture
The NHS of women using media while eating was significantly lower than the NHS of women using no media (β = –0·18 (95 % CI –0·27, –0·09), P < 0·001). Furthermore, women usually snacking between meals had a significantly lower NHS than women not snacking between meals (β = –0·24 (95 % CI –0·33, –0·15), P < 0·001). The number of meals was not significantly associated with the NHS (β = 0·08 (95 % CI –0·01, 0·17), P = 0·078).
Discussion
We developed a short questionnaire for evaluating the diet composition and eating culture in pregnant women.
Healthiness of diet reported in the present sample
Our results showed that food items were consumed according to the recommendations (‘healthy’) or only slightly different from recommendations (‘moderately healthy’) which indicate healthy eating in our sample. Pregnant women paid particular attention to an adequate intake of healthy foods (e.g. fruits, vegetables, unsweetened milk products) and avoided unhealthy foods (e.g. ready-made meals). The transition to motherhood is associated with changes in eating behaviour(Reference Olson27–Reference Szwajcer, Hiddink and Koelen29) and increased consumption of fruits and milk products during pregnancy(Reference Forbes, Graham and Berglund28). The avoidance of specific food items (e.g. (raw) meat, raw fish) might be due to women’s concerns about child health, aversion to different food items and nausea during pregnancy(Reference Forbes, Graham and Berglund28). However, nearly 16 % of women reported that they eat raw products during pregnancy, contrary to recommendations(Reference Koletzko, Bauer and Bung1). Nearly 7 % of pregnant women followed a vegetarian or vegan diet. Unfortunately, we cannot say whether these (or even more) women had followed a vegetarian or vegan diet before pregnancy.
In our analysis, sweetened beverages and treats were, on average, consumed more frequently than once a day (8·28 portions/week and 9·48 portions/week, respectively) during pregnancy. Although it has been shown earlier that sweetened drinks are consumed frequently in other population groups and are further associated with overweight and obesity – due to underestimation of energetic intake – we suggest that during pregnancy, increased intake of sweetened drinks and treats might also be due to pronounced preferences for sweet taste during pregnancy(Reference Bowen30).
Validity: comparison with FFQ
The answers given in CoCu pregnancy showed significant positive associations with answers given in the FFQ. The correlations for sweetened milk and white bread were significant, but weak, which confirms our observations of CoCu in children and adolescents(Reference Poulain, Spielau and Vogel20).
For both sweetened milk products and white bread, the weak correlations might be explained by the difficulty to correctly estimate the daily use, e.g. due to irregularities between different days (in the case these products are not consumed every day) or due to difficulties in distinguishing different types of food (e.g. yoghurts containing sugar or not; bread containing grains or not). For sweetened milk products, methodological differences between the questionnaires might also be responsible for the weak correlations. For example, fruit yoghurt is categorised as unsweetened milk product in the FFQ, but as sweetened milk product in the CoCu.
Overall, we were able to confirm the validity of all food items in CoCu pregnancy.
Agreement
We assessed percentage agreement of CoCu pregnancy in a subsample of ninety-seven women who had completed the questionnaire at two time points during pregnancy (24th and 36th wp). Our results demonstrate that responses for food consumption during pregnancy varied between study visits but that the large majority of pregnant women had chosen at least adjacent answer categories. The results indicate that eating habits in mid- and late-pregnancy are different and changes in diet during pregnancy are common. Our results indicate that eating habits for sweet and savoury snacks remain consistent in only 35 % of subsample which is in line with a previous study demonstrating changing sweet appetite during pregnancy(Reference Belzer, Smulian and Lu31).
The high degree of agreement among the answers on fish consumption between 24th and 36th wp (71 %) might be explained by overall small consumption of fish (one portion per week or less). Consequently, variations between answers remain small.
Associations between diet composition/raw products/eating culture and age/socio-economic status
We further investigated whether diet composition, the consumption of raw products and eating culture are associated with age or SES in pregnant women.
Our results indicate that a healthy diet, a higher number of meals per day and an infrequent use of media while eating are associated with increasing age in pregnant women. We suggest that less health awareness is more prevalent in younger women and increases during adulthood. A study investigating the adherence to the Mediterranean diet and their association with different maternal factors during pregnancy verified better adherence to the Mediterranean diet with increasing age of pregnant women(Reference Jardi, Aparicio and Bedmar19). Results from the NICE study indicated that increasing age was associated with a diet rich in fruits, vegetables, whole grains and fish while pregnancy(Reference Stråvik, Jonsson and Hartvigsson32). The increased health awareness is also reflected in the fact that in our cohort, meal frequency was associated with age. While regular meals are recommended for children(Reference Koletzko and Toschke33) and adults, skipping meals is associated with adverse health effects, e.g. overweight or obesity(Reference Cho, Dietrich and Brown34).
The higher media use while eating in younger as compared with older women might be explained by a generally greater ubiquity and naturalness of media, especially modern media such as mobile phones, among younger people and parents(Reference Feierabend, Plankenhorn and Rathgeb35).
Furthermore, increasing age of pregnant women might indicate that they already have children and fulfil an enhanced role model function for them.
Our results demonstrated that healthy diet (NHS), less consumption of raw products and unhealthy snacks between meals, higher number of meals per day and less use of media while eating are associated with higher SES in pregnant women. This is in line with other studies in different countries showing socio-economic disparities for diet quality during pregnancy(Reference Jardi, Aparicio and Bedmar19,Reference Stråvik, Jonsson and Hartvigsson32,Reference Li, Lei and Zhang36,Reference Saunders, Rehbinder and Carlsen37) . Additionally, another study demonstrated that pregnant women with lower SES was less compliant in avoiding raw food products(Reference Moran-Lev, Bauer and Farhi38). This is in line with our findings indicating lower consumption of raw products in pregnant women with high compared with middle SES.
People and parents from lower social strata generally show a higher media use than people from higher social strata, especially regarding entertainment media such as video games or television(Reference Feierabend, Plankenhorn and Rathgeb35).
We suggest that higher SES and increasing age are associated with increased health awareness, avoidance of risks, increased knowledge about healthy eating and increased consideration of recommendations applicable during pregnancy(Reference Koletzko, Bauer and Bung1). Within our cohort, socio-economic status seems to be even more important for diet and eating culture than age.
Our results showed that a healthy diet during pregnancy is affected by, e.g. SES and age. Therefore, we recommend practical support such as information on quick recipes, the provision of inexpensive or even free fruit and vegetable products and the promotion of subsidised, delivered lunches for pregnant women.
Associations between diet composition and eating culture
In our analysis, pregnant women with a high NHS used less media during meals. This confirms our results from analyses in children and adolescents(Reference Poulain, Spielau and Vogel20) and was described in previous studies(Reference Cha, Hoelscher and Ranjit39,Reference Ellithorpe, Eden and Hahn40) . In our cohort, snacking was associated with lower NHS in pregnant women. Snacking has been further identified as potential risk factor for poor food quality in children(Reference Taillie, Afeiche and Eldridge41), adolescents(Reference Larson, Miller and Watts42) and adults(Reference Murakami and Livingstone43). Our results indicate that pregnant women who are concerned about their diet composition also maintain an appropriate eating culture.
Implications
We created and validated the CoCu pregnancy as a useful and time-efficient tool assessing diet composition and eating culture in pregnant women. To the best of our knowledge, no other questionnaire combines questions on diet composition and eating culture in pregnant women.
The questionnaire might be used in epidemiological studies as well as in clinical practice. In consultation hours, information about deviations from recommendations during pregnancy could be identified more quickly and might help making further dietary advices more easily.
Limitations
We acknowledge that, like in other cohort studies, an increased number of participating women showed a high or medium SES. Therefore, it is possible that findings would differ in pregnant women from lower social strata. Furthermore, the number of questionnaires included into the analysis was rather small.
Another limitation is that the assessment of diet is subjective and we cannot exclude responses based on social desirability. With regard to the validity check, we were able to show a positive significant correlation for all food items, of those two food items showed only a low correlation. To validate the questionnaire, the comparison was made with only one other instrument (FFQ). This was due to other extensive study investigations. The output of, e.g. food diary or 24-h-recalls, would have exceeded the time frame for each participant.
Finally, the LIFE Child study does not assess diet consumption and eating culture before pregnancy. Therefore, we could not compare diet before and during pregnancy.
Conclusion
The created short nutrition questionnaire represents an economic and efficient tool for assessing the diet composition and eating culture in pregnant women and might be used in epidemiological studies as well as in clinical practice. The results demonstrated that both healthier diet composition and eating culture were associated with increased age and higher SES of pregnant women.
Furthermore, findings from the questionnaire can be used in the future to ensure better care for mother and child in the context of prevention.
Acknowledgements
Acknowledgements: We thank all women for their participation. We acknowledge support from Leipzig University for Open Access Publishing. Financial support: This publication is supported by LIFE – Leipzig Research Center for Civilization Diseases, University of Leipzig. LIFE is funded by means of the European Union, by means of the European Social Fund (ESF), by the European Regional Development Fund, and by means of the Free State of Saxony within the framework of the excellence initiative. U.S. was supported by the Federal Ministry of Education and Research, Germany (grant number FKZ01EO1001 and FKZ01EA1908B). The funding institutions had no role in the design, analysis or writing of this article. Conflict of interest: None of the authors had a conflict of interest. Authorship: T.P. and U.S. created the questionnaire were responsible for conception and design of the study, analysed and interpreted the data and were major contributors in the writing of the manuscript; M.V. contributed to the analysis of data; A.K. and W.K. supervised conception and design of the study, data acquisition, analysis and interpretation, and contributed to the writing of the manuscript. All authors drafted the article and gave final approval to the version to be submitted. Ethics of human subject participation: The LIFE Child study was designed in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University of Leipzig (Reg. No. 264-10-19042010). Informed written consent was obtained from all participants.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980021002445
