Food Consumption, Nutrient Intake and Status during the First 1000 Days of Life in The Netherlands: A Systematic Review

Adequate nutrition is essential for growth and development in early life. Nutritional data serves as a basis for national nutritional guidelines and policies. Currently, there is no insight into the availability of such data during the first 1000 days of life. Therefore, a systematic review was performed, following the PRISMA reporting guideline, to identify studies on food consumption, nutrient intake or status in the Netherlands. Potential gaps were identified, and the quality of the studies is discussed. The databases Embase and Medline were used, as well as databases from national institutes. Articles published in 2008–2018 were screened by two independent reviewers. In total 601 articles were identified, of which 173 were included. For pregnant women, 32 studies were available with nutritional data, for young children 40 studies were identified. No studies were available for breastfeeding women. A large variety of foods and nutrients were assessed, however certain nutrients were lacking (e.g., vitamin K). Overall, the studies had methodological limitations, making the data unsuitable to assess nutrient inadequacies. There is a need for recent, high quality nutritional research to strengthen the understanding of the nutritional needs and deficiencies during early life, and is fundamental for national guidelines and policies.


Introduction
Adequate nutrition is essential for growth and development in early life [1]. Sufficient energy and nutrient intakes during pregnancy are needed to support maternal tissue growth, as well as to prevent unfavorable birth outcomes. A well-known example is the relation between folic acid intake of the mother and neural tube defects in the offspring [2]. In addition to these relative short-term effects, there are long-term effects of prenatal nutrient exposure. Evidence is growing that early exposure, during pregnancy or even periconceptionally, can affect the metabolism of the child. It is suggested that maternal pre-pregnancy obesity is associated with offspring adiposity, potentially through leptin epigenetic regulations [3,4]. Results from the Dutch famine birth cohort study show an association between severe prenatal malnutrition and the development of obesity and coronary heart disease later in life [5]. This relation between the metabolic programming in early life, and the development of non-communicable diseases in later life, is known as the 'Developmental Origins of Health and Disease (DOHaD) hypothesis' [6][7][8].
Although there is little concern for any energy insufficiency in affluent countries, pregnant and breastfeeding women may be at risk of certain nutrient deficiencies, due to an increased nutritional need, and low nutrient-dense food patterns. Previous research indicates that pregnant women, living in affluent countries, are at risk of folate, iodine, iron and vitamin D inadequacies [9,10].
A PICO model was used to formulate the search strategy [25]. The population (P) was defined as: First 1000 days, pregnant women, mothers during the breastfeeding period and children up to two years of age. Populations living in the Netherlands without medical illnesses were included. Intervention (I) studies were excluded, except when baseline data was available, prior to the intervention. A comparison to a control group (C) was not taken into account, as we did not study an intervention. However, if nutritional data was available from a healthy control group, these data were included. The outcomes (O) of interest were data on food and nutrient intake, dietary supplement use and biochemical nutrient intake markers. Emtree index terms were used and exploded to find as many relevant studies as possible. The search filters: Humans, publication data and publication type, were used. No language restrictions were applied. The search string is provided in the Appendix A Table A1. In addition, reports from the National institute of food safety (RIKILT), the National Institute for Public Health and the Environment (RIVM) and the Netherlands Organization for applied scientific research (TNO) were identified through the institute-websites [26][27][28]. For the identification and selection of relevant reports, the same PICO and exclusion criteria were used as for the scientific articles.
Titles and abstracts were screened independently by two reviewers (N.K., S.t.B.). Predefined exclusion criteria were used: Studies published before 2008; did not contain Dutch data; included a population with a medical illness or premature infants; the population was not pregnant, breastfeeding, or had a mean age above two years; preconception data; no data on food or nutrient intake, supplement use or nutrient status; intervention studies without reporting baseline data; case studies. Subsequently full texts were retrieved and assessed based on the selection criteria stated above. If the full text could not be retrieved, they were excluded. Disagreement on the inclusion of a study was resolved through consensus or the consultation of a third reviewer (J.V.-K.). The flow diagram of the study selection is provided in Figure 1. For each of the included studies the following data was extracted by the two reviewers (NK, StB): Study name, year of publication and years of data collection, type of study (cohort, case-control, randomized controlled trial), study population characteristics (i.e., location, gestational age and birth weight, age, ethnicity, BMI), supplement intake data, dietary assessment method and validation, reported foods, nutrients and biochemical markers of nutrient intake status. Data on breastfeeding was considered as a food intake of children, although the assessment took place among their mothers. For data interpretation, articles referring to the same study or cohort name were grouped.
The reference management software EndNote X9 was used during the selection procedure. The study characteristics were recorded in Microsoft Excel (Microsoft Office Professional Plus 2010).

Results
In total 601 articles were identified, of which 173 met the inclusion criteria ( Figure 1). The main reasons for exclusion were the absence of nutritional data, and a non-relevant study population. Of the 173 articles, 109 articles were identified with nutritional data on pregnant women, zero articles with data on breastfeeding women, and 114 articles with data on children up to the age of two years. Certain articles contained information on both pregnant women and young children. The study characteristics can be found in Table 1 for pregnant women and Table 2 for children up to two years of age respectively. The cohorts with the largest number of publications were the Generation R study (n = 78), the ABCD study (n = 17), KOALA study (n = 7) and PIAMA study (n = 6). Publications with data on pregnant women originated from 32 different studies. For children up to two years of age, 40 different studies were identified. The number of study participants ranged from 8 to 7857 pregnant women, and 9 to 7857 children.

Food Consumption
For pregnant women, the most frequently reported data, on food consumption, were on alcoholic beverages (n = 18), fish (n = 5), fruit, and vegetable consumption (both n = 4) ( Table 3). Alcohol exposure (yes/no) during pregnancy was assessed, and to a lesser extend the amount which was consumed. Although there was a large variety in the foods that were assessed, these originated mainly from one single cohort (i.e., Generation R study). The most frequently reported data for children up to two years of age was on feeding practices (i.e., breastfeeding, formula feeding) ( Table 3). Whether children were breastfed was reported in 28 different studies, and the period of breastfeeding was also frequently assessed. Formula feeding was reported in 12 different studies.

Nutrient Intake
Maternal energy and macronutrient intake were most frequently stated, in up to five different studies (Table 4). For the vitamins, data was reported for retinol, riboflavin, nicotinamide and folate. These data originated from one single cohort (i.e., Generation R). No data was available for beta-carotene, thiamine, pantothenic acid, pyridoxine, biotin, cobalamin and vitamins C, D, E and K. For the maternal mineral intake, data was available for calcium, iron, magnesium, phosphorus and sodium. As with vitamins, the mineral intake data originated from one single cohort (i.e., Generation R). No data was available for chlorine, chromium, copper, iodine, manganese, molybdenum, potassium, selenium and zinc intake. Two studies reported on caffeine intake (i.e., Generation R and ABCD study). For children up to two years, primarily energy and the macronutrients were reported, by four and three different studies, respectively. A large variety of vitamins and minerals were assessed, however originating from one or two different studies (i.e., Generation R and Eat complete test). Data was reported for: Thiamine, riboflavin, niacin, vitamin A, beta-carotene, pyridoxine, cobalamin, folate, vitamins C, D and E, calcium, copper, iodine, iron, magnesium, phosphorus, potassium, selenium, sodium and zinc. In addition, data was available on water intake. No data was available for pantothenic acid, biotin, vitamin K, chlorine, chromium, manganese and molybdenum.

Biochemical Nutrient Status
Most frequently stated maternal nutrient status markers were fatty acids and vitamin D, which were reported by six and four different studies, respectively ( Table 5). The B vitamins (folate, cobalamin) and homocysteine were stated by two to three studies. Retinol, ferritin, iodine, and iron were reported in one study each. No data for vitamin C, calcium, copper, selenium and zinc status was available. For children up to two years of age, nutrient status data was available for vitamin D (three different studies), fatty acids (one study) and ferritin (three different studies) ( Table 5). For calcium, one study provided the data. No data was available for the B vitamins, vitamin C, copper, iodine, iron, selenium and zinc status.

Supplement Use
The most frequently reported supplement was folic acid, being about whether women used this supplement during their pregnancy (thirteen different studies, Table 6). Several studies reported on the duration and the period (e.g., preconception) of folic acid supplement use. For children, the most often assessed supplement was vitamin D, which was reported by ten different studies (Table 6).

Time Span
The years in which the data were collected ranged from 1989 until 2016 (see Tables 3-6). For pregnant women, the most recent data originated from 2016 (alcohol consumption), however most data was collected before 2006. Nutrient status data was more recent, with assessments of vitamin B 12 , folate and homocysteine up to 2014. For children up to two years of age, most food consumption and nutrient intake data were collected before 2014. The most recent nutrient status data originated from 2014 (ferritin), the other status markers (such as vitamin D) were assessed before 2009.

Dietary Assessment Methods
For the assessment of alcohol use, breastfeeding practices and folic acid supplement use, general questionnaires were used. Food consumption and nutrient intake were mainly assessed via a semi-quantitative food frequency questionnaire (FFQ) (see Tables 3 and 4). In addition, a two day food record was used by five different studies, and a dietary recall was performed by three studies (of which one performed a repeated dietary recall). There was one study which used a semi-weighted food record, and one study which used duplicate portions. Often no information on any validation of the methodology was provided. In case the method was validated, it was generally used after adaptations, and not in its original form. Only one study explicitly reported the validation of the FFQ among pregnant women.

Discussion
To our knowledge, this review is the first to provide a comprehensive overview on Dutch nutritional data, and to include the full population-range of the first 1000 days of life. In addition, nutritional intake data as well as biochemical nutrient status were included in this review, which are essential for determining potential nutrient deficiencies or excesses.

Previous Findings
Previously, a systematic review was published on nutrient intake and biochemical nutrient status among pregnant women in affluent countries [9], and more recently on pregnant adolescents [202]. The first review however included one Dutch study, performed in 1988. The second review, on pregnant adolescents, didn't include data from the Netherlands. Although we identified data for pregnant women and young children, no data was found for breastfeeding women living in the Netherlands. There are however a few studies on breastfeeding women in affluent countries: A recent study assessed the dietary intake among French breastfeeding women, reporting their energy and macronutrient intake [203], and the energy and macronutrient of Greek breastfeeding women was assessed [204]. Other studies focus specifically on vitamin D intake and status [205,206]. Additional research is of interest, as adequate nutrition during this period within the 1st 1000 days (i.e., the breastfeeding period) contributes to maternal and infant health [207].

Strengths and Limitations
The present review provides a detailed overview of the available studies with nutritional data, there are however certain limitations that need to be addressed. Although the search string was developed with great care, we cannot guarantee that all relevant articles were identified. Unpublished data and publication bias may influence the findings, as well as the choice of search terms; however, as a large number of articles were identified, it is expected that our conclusions are robust. We identified multiple publications which referred to the same study. To prevent an overestimation of the available data, we grouped these references, assuming that these articles refer to an identical population. This assumption may not be correct for all articles, as subpopulations of a study may have been assessed, and published in separate articles. Studies which did not report a cohort or study name were assumed to be independent studies. These assumptions may affect the number of individual studies, which were available for the nutritional parameters.

Quality of the Studies
Although the results indicate that there is nutritional data available, the usability of the data, to evaluate the nutrient intake of pregnant women and children up to two years of age, is limited. Most of the identified data was from studies examining the association between maternal life styles, environmental determinants and health outcomes. Only a few studies were designed to evaluate the nutrient adequacy. Factors such as the time span, national representativeness and methodology affect the usability of the data, and will be discussed below.
This review includes data published since 2008. Although the most recent data originates from 2016, many of the assessments were performed prior to 2006. It can be questioned as to whether these data are still representative for the current dietary intake, as dietary patterns and food compositions may change over time [208]. In addition, changing national legislations and policies may have affected the nutrient composition of the food: In the Netherlands, fortification of foods with vitamin D and folic acid are allowed since 2007 [209]. In 2008 the Dutch policy on the addition of iodized salt changed, resulting in a decrease of iodine intake among the Dutch population [210].
Although the identified studies are dated, there are several cohort studies, which are currently running in the Netherlands. These may provide future data on pregnant women and their offspring. We are however unaware of ongoing studies among lactating women.
Most of the identified studies were not nationally representative for pregnant women or young children living in the Netherlands. Studies focused for instance on specific cities or areas within the Netherlands. The National Food Consumption Survey [144] contained recent and representative data (2012-2014) on young children (aged 1-3 years). Pregnant women and breastfeeding women were however excluded from this survey.
A variety of dietary assessments were used. Most often articles stated that a general questionnaire was used, without further specifications. These questionnaires were utilised to give an impression of alcohol and dietary supplement use during the pregnancy. In addition, these questionnaires were employed to assess exclusive breastfeeding and the duration of breastfeeding. For dietary assessments, the food frequency questionnaire was most often used. FFQs are a cost-effective and time-saving method, with a low participant burden, which make them suitable for large epidemiological studies [211]. Although this method can be used to rank persons based on their intake, it often does not provide a valid estimate of the actual consumed amount, or the distribution of the (habitual) intake of the population, due to the generally limited number of items included in an FFQ [212,213]. This habitual intake distribution is needed to estimate potential deficiencies and excess intakes at population level, and can be obtained by repeated detailed data collection, e.g., dietary records or 24 h recalls corrected for the within-person variation. Although these methods have their own limitations, they perform better with respect to estimating the nutrient intake [214,215]. A few studies were identified which used two day dietary records, however these assessments were prior to 2006, or not in a national representative sample. Almost all studies relied on self-reported dietary intakes, which may influence the results. Only two studies used a (semi-) weighted assessment method.
Validation was often absent or poorly described in articles, and in some cases, the validation was performed in an inappropriate study population, such as older adults. One study mentioned that they used an FFQ that was validated in pregnant women [129]. The authors however recommend further validation, as they adjusted the questionnaire based on a pre-test by introducing plate photos.
Not all studies specified the nutrient database which was used for calculating the nutrient quantities. As these databases change over time, using a dated version may influence data quality.

Absence of Data for Certain Nutrients
For some nutrients, the intake or status was not assessed or it was scarce. This may be due to the study aim, focused on a single specific nutrient, or due to methodological constrains. For nutrient status, the absence of a good status marker, feasibility of the assessment method, or high costs, may be restricting factors, and might explain why certain data are scarce.
Serum calcium for instance, is highly regulated within the body, and is strongly affected by other nutrients such as vitamin D, and so it has limited use for assessing calcium status [216]. This may explain why there was only one study found which assessed calcium status. Another example is iodine. Urinary iodine is a useful biomarker to assess the iodine intake status [217], it has however multiple disadvantages: A high burden on the study participants, high costs and possible incomplete 24 h urine collections. These factors may explain why it is not frequently included in studies. Our results identified only one study assessing iodine status, via a single spot urine sample, among pregnant women. Spot urine collection is less burdensome, however additional research is needed on its validation in pregnant women and children [218]. Spot urine uses an assumption regarding the 24 h urine volume, this assumption may however be incorrect, as a higher urinary volume was measured in the Dutch adult population [219].

Suggestions for Future Research
Overall, there is a lack of recent, representative nutritional data for pregnant women and young children. For some nutrients, and for breastfeeding women, the data were absent. There is a need for additional research: A future study should include a national representative population, and research is needed on breastfeeding women. In addition, in order to assess nutrient adequacy, the study should include a validated dietary assessment, such as a repeated 24-h dietary recall, combined with certain biochemical status assessments, especially for nutrients difficult to assess with, e.g., food consumption surveys. Examples are vitamin D, which is strongly influenced by sunlight exposure, and iodine, which is provided by fortified salt and is difficult to quantify through dietary assessment methods.

Conclusions
The current systematic review provides a comprehensive overview of the available nutritional data for the first 1000 days, assessed in the Netherlands. Although there was a large variety of foods and nutrients which were reported, most originated from one single study. No nutritional data was found for breastfeeding women, and the majority of assessments were performed before 2006. This time span, the methodology used and the absence of data for certain nutrients, make it difficult to evaluate the current nutritional intake and status of pregnant women, breastfeeding women and children up to two years of age. Overall, there is a need for high quality nutritional research during the full period of the first 1000 days of life. This data will strengthen the understanding of the nutritional needs and the potential nutrient deficiencies during early life, and is fundamental for national guidelines and policies.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.