Adherence to the Norwegian dietary recommendations in a multi‐ethnic pregnant population prior to being diagnosed with gestational diabetes mellitus

Abstract Maternal diet is a modifiable risk factor for the development of gestational diabetes mellitus (GDM). Even though pregnant women are considered to be motivated to eat healthy, previous research found unhealthy eating patterns among some ethnic and lower socio‐economic status groups. This cross‐sectional study assessed adherence to national dietary recommendations prior to GDM diagnosis in a multi‐ethnic population comprising 237 pregnant women. Participants were diagnosed with GDM after performing a two‐hour oral glucose tolerance test ≥ 9 mmol/L. Participants answered a 41‐item Food Frequency Questionnaire about dietary habits prior to being diagnosed with GDM from October 2015 to March 2018. Their scores were based on adherence to the recommended intake in each food group and summed into a Healthy Diet Score (HDS). Results showed low adherence to national dietary recommendations. A significantly higher proportion of non‐native Norwegian‐speaking women had a high HDS compared with native Norwegian‐speaking women. Participants with a normal prepregnancy weight were more likely to have a high HDS compared with overweight or obese participants. Participants showed low adherence to the recommendations for whole grains, vegetables, and fruits and berries, and a relatively low proportion adhered to the recommendations for intakes of fish, red/processed meat, and ready‐made meals. However, the food group intakes varied by country of birth. Given the increase in women with GDM and the emerging evidence that maternal diet is a modifiable risk factor for GDM, effective nutrition communication strategies in antenatal care are urgently needed.

type 2 diabetes mellitus (T2DM) and cardiovascular disease for the mother and child (Shah, Retnakaran, & Booth, 2008). For instance, women who have had GDM are, at least, at a seven-fold increased risk of developing T2DM compared to those with euglycemic pregnancies (Bellamy, Casas, Hingorani, & Williams, 2009). Furthermore, children born to mothers with GDM already face a higher risk of developing T2DM at a younger age and obesity and cardiovascular disease later in life (Gluckman, Hanson, & Buklijas, 2010). Thus, efforts to prevent GDM are paramount.
The prevalence of GDM is increasing worldwide and ranges from 1% to 20% globally depending on the screening procedure and population characteristics (Rani & Begum, 2016). The prevalence is increasing in Norway as well and was estimated at 5.2% in 2016 (Medical Birth Registry Norway, 2018). However, a cohort study in Groruddalen, a district in Oslo, identified GDM in 13% of all women, 11% in ethnic Norwegians, and 12%-17% in groups of non-European origin (Jenum et al., 2012). Women from South Asian and African backgrounds tend to develop GDM at a lower BMI and age compared with white Europeans (Makgoba, Savvidou, & Steer, 2012). Other risk factors for developing GDM include overweight and obesity, advanced maternal age, a family history of diabetes, and GDM in a previous pregnancy (Schneider et al., 2011). As a large population-based cohort study in the Netherlands found, when adjusting for ethnicity, family history of diabetes, and parity, low-educated women faced an increased risk of GDM mainly due to higher rates of overweight and obesity (Bouthoorn et al., 2015).
Dietary intervention studies on the prevention of GDM are abundant but there is, as yet, no conclusive evidence on what constitutes the optimal dietary pattern for the prevention of GDM (Olmedo-Requena et al., 2019;Schoenaker, Mishra, Callaway, & Soedamah-Muthu, 2016;Shepherd et al., 2017). In the Norwegian guidelines for antenatal care, health professionals are encouraged to provide women with information about a healthy, varied diet in line with the national dietary recommendations in order to prevent foodborne diseases (Directorate for Health & Social Affairs, 2005).
In particular, women are encouraged to eat whole-grain products, fruits and vegetables, lean dairy, and meat products and to limit their sugar and salt intakes. However, studies indicate that pregnant women received little nutrition-related advice in antenatal care, and participants from immigrant backgrounds appeared to be confused about dietary advice that was incongruent with their original food culture (Garnweidner, Sverre Pettersen, & Mosdol, 2013;Szwajcer, Hiddink, Koelen, & van Woerkum, 2005).
Although pregnant women are considered to be motivated to eat healthy, several studies in different countries found unhealthy eating patterns among pregnant women (Shapiro et al., 2016;Zhu et al., 2018;von Ruesten et al., 2014). According to a contemporary multiracial prospective cohort study, the majority (79%) of pregnant women did not adhere to the Dietary Guidelines for Americans (Zhu et al., 2018).
Moreover, ethnic differences in maternal dietary patterns were found (Sommer et al., 2013). A cross-sectional study among 757 pregnant women in Norway found that all non-European women, as compared to Europeans, were more likely to have the unhealthier dietary pattern (Sommer et al., 2013). Given the increased risk of GDM in women of African and South Asian origin, it is important to gain more knowledge about possible ethnicity-related differences in the dietary quality of pregnant women.
Thus, the aim of this study was to assess adherence to national dietary recommendations in a multi-ethnic pregnant population prior to being diagnosed with GDM.

| Design and study sample
A cross-sectional study was performed using baseline data from the Pregnant+ study, a randomized controlled trial (RCT) among preg- Data were collected from 237 pregnant women, all of whom were diagnosed with GDM after performing a two-hour oral glucose tolerance test (OGTT) ≥ 9 mmol/L. The period between when participants received the diagnosis and filled out the questionnaire varied from one to seven days. The definition of GDM was in accordance with the national guidelines for antenatal care and the WHO (Legeforeningen, 2008;World Health Organization, 2013). To be included in the study, women had to have a smartphone, be 18 years or older, and be at a gestational age of at least 33 weeks. The women had to be capable of filling out the questionnaire in either Norwegian, Somali, or Urdu.
Only 14 women filled out the questionnaire in either Urdu or Somali.
Participants were excluded from the study if they had a twin pregnancy. In addition, women with celiac disease or lactose intolerance were excluded because they must follow special diets (Borgen et al., 2017). The study was approved by the Norwegian Social Science Data Services (ID number 2014/38942), and the patient privacy protections boards governing each of the recruiting sites. Written consent was obtained from all participants.

| Measures
Women answered the questionnaires on an electronic tablet at their first consultation at a DOC. Participants were asked to complete a 41-item Food Frequency Questionnaire (FFQ) reporting their dietary habits prior to being diagnosed with GDM. The FFQ included the following food groups: beverages, milk and dairy products, bread and grain, fruit and vegetables, snacks, meat, and processed foods. Answers to the questions on frequency of intake ranged from 0 (never) to 9 (several times daily). The FFQ was based on the Fit for Delivery study and was shown to have an adequate level of test-retest reliability (Øverby, Hillesund, Sagedal, Vistad, & Bere, 2015). The FFQ in Somali and Urdu was pilot tested among Somali and Pakistani Norwegian women. Participants were given scores based on adherence to the recommended frequency of intake, which varied among the different food groups. The adequacy of intake of healthy foods and moderation of intake of less healthy foods were taken into consideration (Brosig, Burggraf, Teuber, & Meier, 2018), and points of 0, 5, or 10 were given for the level of adherence to each food group ( Table 2). The scores were summed into a Healthy Diet Score (HDS), which was categorized into tertiles of "low" (<40 HDS), "medium" (45-60 HDS), and "high" (65-120 HDS) adherence to the recommendations.

| Statistical analysis
Statistical analyses were performed with SPSS for IBM statistical software package version 24.0 (IBM Corporation, Armonk, NY, USA).
Cross-tabulations with Pearson's chi-square tests were used to calculate percentages and assess differences in background characteristics and adherence to a healthy diet by HDS tertiles. Multinomial logistic regression analysis was used to examine the relationship between the different HDS (high, medium, and low) and background characteristics. Univariable models were performed first, with the HDS as the dependent variable. All variables in the preliminary univariable models were included in a multivariable model if they were significantly associated with one or more of the different levels in the HDS. Both language (categorized as non-native and native Norwegian speaking) and country of birth (categorized as Norway, Western Europe and United States; Eastern Europe; Asia; Africa; and South America) were significantly associated with the HDS. Country of birth was retained in the final model to show more nuanced differences. A p-value below .05 was the level of inclusion.
Differences in frequency of food group intakes by categories of BMI and by country of birth were analyzed via an analysis of variance (ANOVA). Differences in intakes between two countries/regions were analyzed via a Student t test.

| RE SULTS
A total of 237 women were included in the study. The majority were between 24 and 32 years old (76.8%). Mean gestational age when filling out the questionnaire was 26.7 weeks (SD = 4.9). Mean prepregnancy BMI was 26.7 (SD = 5.7) (not in tables). Table 1 contains background characteristics for the whole sample and according to the tertiles of HDS. A total of 108 (45.6%) participants were native Norwegian speakers, and 129 (54.4%) were non-native Norwegian speakers. Nearly 25% (24.1%) were born in Eastern Europe and Asian countries. A total of 12.2% were born in African countries, while 5.9% were born in Western countries, including the United States, and 2.1% in South American countries. The proportion of women with poor language skills was very small (7.4%).
Participants were categorized as having low (28.7%), medium (36.7%), or high (34.6%) HDS. A significantly higher proportion of non-native Norwegian-speaking women had a high HDS compared with native Norwegian-speaking women (Table 1). Furthermore, compared with overweight or obese women, a significantly higher proportion of participants with normal weight prepregnancy had high HDS. Dietary scores were not associated with women's socioeconomic status and education.
Concerning the intake of food groups included in the HDS, participants showed low adherence to the recommendations for whole grains, vegetables, and fruits and berries (Table 2). A relatively low proportion adhered to the recommendations for intakes of fish, red/ processed meat, and ready-made meals.
When adjusting for country of birth, obese women (BMI 35-45) had a nine-fold increase in the odds ratio of having a low HDS compared with normal weight women (AOR = 9.22, 95%; CI 1.89-47.17) ( When examining food group intake by BMI category, the frequencies of intakes of red/processed meat and ready-made meals increased significantly with increasing BMI category (Table 4).
Furthermore, food group intakes varied by country of birth (Table 5).
Whole-grain foods were significantly more commonly consumed among women born in Norway and Western Europe/United States compared with Asian-born women. Women from Norway and Asia had significantly lower intakes of fruits and berries than women born in Western Europe/United States and Eastern Europe. Fish was most commonly consumed among African-born women, with a significantly more frequent intake than both Norwegian-and Asia-born women. Beans and lentils were most commonly consumed by women born in Africa and South America. Red/processed meat was significantly more commonly consumed among Norwegian-born women compared with women born in Eastern Europe. Norwegian women had the most frequent consumption of ready-made meals and significantly more frequently than women born in Eastern Europe, Asia, and Africa. Sugar was significantly more often consumed by women born in Eastern Europe compared with women born in Norway.   Few studies have investigated ethnic differences in dietary quality of pregnant women (de Seymour et al., 2016;Sommer et al., 2013;Zhu et al., 2018). A cross-sectional study among 757 pregnant women, of whom 59% were of non-Western origin, found that non-European ethnic origin and integration scores were associated with higher odds of having unhealthier dietary patterns (Sommer et al., 2013). In the present study, a significantly higher proportion of non-native Norwegian-speaking women had a high

| LI M ITATI O N S
The results of this study must be interpreted in the context of some  (Waijers, Feskens, & Ocke, 2007). Lastly, there was no assessment of whether participants had received advice regarding a healthy diet from health professionals during antenatal care.

| CON CLUS IONS
This study evinced low adherence to national dietary recommendations among pregnant women. The adherence to national dietary recommendations was significantly lower among native Norwegianspeaking women as compared to non-native Norwegian-speaking women. Although there were different food group intakes related to country of birth, participants had low adherence to dietary recommendations which may prevent the development of GDM. Given the increase in women with GDM and the emerging evidence that maternal diet is a modifiable risk factor for GDM, effective strategies for nutrition communication in antenatal care are urgently needed. According to the findings, these strategies should focus on promoting a healthy diet that is high in whole grains, fish, vegetables, and fruits and berries and low in red and processed meats and ready-made meals.

ACK N OWLED G M ENTS
The authors would like to thank the women who participated in the Pregnancy+ study as well as the health professionals and bachelor and master students involved in the recruitment process.