Dietary habits among men and women in West Greenland: follow-up on the ACCEPT birth cohort

This study is based on the ACCEPT birth cohort [18, 19] established during 2010–2015 in Greenland with the overall aim of exploring environmental exposures, dietary changes, and health during a period of lifestyle transition and climate change. In total, 614 Greenlandic pregnant women were recruited from 16 towns distributed over five regions of Greenland (North, Disko bay, West, South, and East). To be included in the ACCEPT study, the women had to be over 18 years of age at the time of inclusion, lived more than 50% of their lives in Greenland, and have at least one Inuit parent. Of the recruited ACCEPT women, 504 fulfilled the inclusion criteria and 478 completed their pregnancy with available birth outcome (Fig. 1).

During this ACCEPT follow-up study (May 2019 to January 2020), 295 live born singleton children were in the relevant age group of 3–5 years of age (children of mothers recruited in 2013–2015). Of these, 150 fulfilled the follow-up inclusion criteria (mothers had lived longest in the ACCEPT regions West or Disko Bay and currently lived in Nuuk, Sisimiut, or Ilulissat). The child’s biological father was included in this follow-up if possible. We contacted 133 ACCEPT mothers and her partner/child father, and 102 agreed to participate (one family participated with two children from two independent pregnancies). Finally, the total study population at follow-up included 101 mothers, 76 fathers, and 102 ACCEPT children (Fig. 1). The participation rate at follow-up was 76.6%, and those who did not accept to participate mainly gave lack of time as the reason.

After receiving a detailed description of the study, all participants gave signed informed consent to participate. They were informed that they could withdraw their consent at any time in the process and that their participation was voluntary.

The families got two home visits from a health nurse visitor and project researchers, respectively. At the first visit, the health nurse visitor interviewed the families, and questionnaires were filled out, and at the second visit, biological samples (blood, hair, urine, and nails) were collected.

The study was carried out in accordance with the Helsinki convention II, and was approved by the Commission for Scientific Investigations in Greenland (KVUG 2019–04).

The adult participants completed a self-administered questionnaire in Danish or Greenlandic. The health nurse visitors were available for assistance if the participants were in doubt about the meaning of the questions and possible answers. Two independent person’s double-entered data from the questionnaire into the validation program EpiData, and any disagreements were discussed.

The adult participants’ questionnaire was divided into two sections; section 1 contained questions about demographics, lifestyle, and health, and section 2 was a food frequency questionnaire (FFQ) about food intake for the last 12 months.

Data extracted from section 1 included age, history of living places (in and outside of Greenland), ethnicity, educational level, income, alcohol intake, smoking history, drug use, body mass index (BMI) (from self-reported height and weight), and number of children.

The dietary intake was calculated based on an FFQ (section 2 of the questionnaire), which was a slightly optimized version of the questionnaire used at inclusion at pregnancy during 2013–2015 [19]. It contained questions about 42 traditional and 23 imported food items, and it was possible to report the consumption frequencies in eight categories ranging from “never” to “several times a day” for each food item. The food items were divided into seven main traditional food groups: Marine mammals, Seabirds, Fish, Dried fish, Shellfish, Terrestrial animals, and Berries and seven main imported food groups: Meat products, Carbohydrate foods, Sauce, Fruit, Vegetables, Fast food, and Sweets and Snacks (Additional file 1: Table S1A). Each food item was given a frequency score representing the time(s) a month the food item was consumed. The food consumption frequencies for the food groups were calculated by summing the scores for the specific food items in the group. See additional file 1 for further explanation and examples of the calculation (Additional file 1: Table S1B, S1C). The frequency score method and similar FFQs have been used previously [18, 19, 21‐23].

Furthermore, the FFQ included a column for seasonal information, with the possibility to answer if the food item was consumed in the spring/summer and/or autumn/winter. However, the percentage of missing data in these season variables was high (up to 60%), and the information was not included in the analyses. Please see Additional file 2: Table S2B.

For continuous variables, Mann-Whitney U test was used to test the difference between groups due to a high degree of non-normal distributed data (assessed by Q-Q plots) for the food variables. For categorical variables, Pearson’s Chi-square test was used to test the difference.

Correlations between intake frequency of food groups and socioeconomic and lifestyle factors were assessed with Spearman correlations.

To assess changes in food intake from inclusion at pregnancy to follow-up for the mother, the individual differences were calculated by subtraction of the reported intake from inclusion at pregnancy from the reported intake at follow-up (follow-up - inclusion). Thus zero (0) reflect equal intake at inclusion and follow-up, negative values reflect higher intake at inclusion during pregnancy, and positive values reflect higher intake at follow-up.

Wilcoxon Signed Rank test (median equal to 0) was used to test parent differences and changes from inclusion at pregnancy to follow-up (for mothers only) in food intake frequency. Furthermore, we assessed the pairwise correlations by calculating Spearman correlations for the food intake frequency. We also calculated partial Spearman correlations by adjusting for different predictors of the food intake frequency. The correlations between the parents were adjusted for the age difference between the parents (continuous), household income (categorical), percentage of life lived in Greenland for both the father and mother (continuous), and highest educational level for both the father and mother (categorical). Whereas, the correlations between inclusion at pregnancy and follow-up of the mother were adjusted for the time from inclusion to follow-up (continuous), change in education level (categorical), change in household income (categorical), change in BMI (continuous), and change in parity (categorical).

Sensitivity analyses by excluding non-Inuit and persons with less than 50% of their life spent in Greenland were conducted.

Table 1presents the characteristics of the participants in the present study (101 mothers/ 76 fathers). The mothers were younger and leaner compared to the fathers. The mothers were also more likely to have a university degree and have never used hash and drugs, but less likely to have a personal income of more than 250.000 DKK per year.

Most of the participants lived in Nuuk (67.2%), while 32.8% lived in the other towns (Sisimiut, Ilulissat, Kangerlussuaq, and Qaqortoq). Due to the follow-up criteria, all mothers had lived longest in either Disko Bay or West, while some fathers had lived longest in other regions (11.4%) or outside Greenland (7.1%). Concerning the ethnic background, 80.0% were Inuit, 17.1% partly-Inuit, and 2.9% non-Inuit. An inclusion criterion for participating in the ACCEPT birth cohort was that the pregnant women were Inuit or partly-Inuit; thus, all non-Inuit participants in the present study were fathers (Table 1).

The results presented here were conducted on the full study population; however, the conducted sensitivity analyses excluding non-Inuit and persons with less than 50% of their life spent in Greenland showed similar results (not shown).

Overall, the median intake frequency for traditional and imported food was distributed as 14 and 86%, respectively (Fig. 2 and Additional file 2, Table S2A). The most frequently consumed traditional food group was marine mammals (median: 6.0 times/month), followed by Greenlandic fish (median: 4.5 times/month) and terrestrial animals (median: 3.5 times/month), while the least frequently consumed traditional food groups were seabirds (median: 1.0 time/month) and berries (median: 1.0 time/month). The most consumed imported food group was carbohydrate food (median: 39.0 times/month) followed by sweets and snacks (median: 28.5 times/month), while fast food (median: 4.0 times/month) was the least consumed imported food group (Fig. 2 and Additional file 2: Table S2A). In Additional file 2 (Table S2B), the intake of the individual food items together with seasonal information can be seen. Of the traditional individual food items, caribou (wild-living Arctic reindeer) was most frequently consumed (median: 2.5 times/month), and 14.5% of the participants ate it at least once a week. For the imported food items, potato, pasta, rice, sauce, vegetables, and fresh fruit were all consumed 13.0 times/month (median), and 79.5–86.4% ate the food items at least once a week.

The intake frequency of traditional food groups was similar for mothers and fathers (Table 2); however, the percentage of traditional food was higher among the fathers (15%) compared to the mothers (13%). A similar difference between men and women has previously been seen in Greenland [24], as well as among other Indigenous populations, including the Sami population in Northern Norway [27] and the Eeyou Istchee communities in northern Quebec [28].

Mothers had significantly more frequent intake of fruit and vegetables and less frequent fast food intake (Table 2). These results are in line with previous observations, both in Greenland and internationally. Generally, studies find that women have a healthier lifestyle and diet [29, 30]. Among the Sami population in Northern Norway, fruit and vegetable intake positively correlated with the female gender [27]. In Greenland, women were more likely to comply with the national dietary guidelines [17], and the mean energy intake from fast food was higher for men (1850 kJ/day) than for women (1485 kJ/day) [24].

Among parents living together, intake from the food groups, both traditional and imported, was positively correlated (Table 3). However, mothers had a more frequent intake of vegetables and fruit and less frequent intake of fast food than the fathers, similarly to what was seen in the gender analyses (Table 2). The positive correlations are not surprising, as the parents share household larders and mostly consume main meals together. Other studies have found that couples tend to eat similar foods and have similar nutrient intakes [31‐37]; however, female partners were often restricted in dinner choices by their male partners’ preferences for meat, fewer vegetables, and little food variety [38]. In the present study, the mothers reported a more healthy diet with more frequent intake of fruit and vegetables and less fast food intake than the fathers. The result may indicate that even though the parents most often eat together, the mothers more often eat fruit and vegetables as side dishes, whereas the fathers more often visit fast food places, for instance, at lunchtime.

The living place can affect the hunting possibility and the availability of traditional foods, and the degree of urbanization also influences the availability and supply of imported food. In the present study, terrestrial animal intake was lower in Nuuk than in the other towns (Sisimiut and Ilulissat), while vegetables and fruit were more frequent consumed in Nuuk. The percentage of traditional food was also higher among participants from the other towns (16%) than Nuuk (13%). The availability of the food items may very well explain these differences. Previously, the Population Health Surveys have reported that participants living in settlements were more likely to eat a meal from their own hunt/catch once a week (80%) than participants from towns (39%) and the main city Nuuk (20%) [24]. In compliance with the present study, in the study on Health Behavior in School-aged Children, the proportion of children eating fruit and vegetables daily was higher in Nuuk than in other towns and settlements [39].

In the ACCEPT cohort, geographical differences were investigated based on the region where the women had lived the longest and not by the current living place [18, 19]. Differences among regions were seen for terrestrial animals, berries, sauce, and fast food. Data from 2010 to 2011 showed a less frequent intake of terrestrial animals and berries in Disko Bay than in West, North, South, and East [18], but this difference was not seen during 2013–2015 [19] and neither in the present study. However, for terrestrial animals, we found that the participants living in Ilulissat (Disko Bay) did have a non-significantly less frequent intake (3.0 times/month) than participants living in Sisimiut (West, 5.0 times/month) and Nuuk (West, 3.5 times/month) (data not shown).

The older age group (≥34.3 years) had a healthier lifestyle than the younger group (< 34.3 years), with more frequent fish and fruit intake and less frequent fast food intake. Similar results have been shown in other studies, internationally [40] and in Greenland [17‐19, 24]. For the young population (18–24 years) in Greenland, the energy consumption from traditional food is lower, the intake of sugar and fast food is higher, and they are less likely to comply with the national dietary guidelines than the older population of ≥60 years of age [17, 24]. Similar to the results in the present follow-up study, at inclusion into the ACCEPT cohort, the younger pregnant women did also have a more frequent fast food intake than the older pregnant women [18, 19]. Thus, the present study is in accordance and confirms previous studies in Greenland.

Participants with high socioeconomic status seem to have a healthier lifestyle than participants with low socioeconomic status. High socioeconomic status (high educational level and high income) positively correlated with more frequent intake of Greenlandic fish, shellfish, dried fish, terrestrial animals, vegetables, and fruit. However, high personal income was also correlated with more frequent intake of sweets and snacks. Other studies have also shown that high educational status tends to correlate with consumption of a better-quality diet [41, 42]. In Greenland, occupation was previously associated with food insecurity (experienced times without any food or money to buy within the last 12 months) among adults, with more unemployed having experienced food insecurity (33%) compared to hunters/fishers (12%) and skilled/trained workers (4%) [16].

We also found that lifestyle factors (BMI, physical/sport activity, alcohol intake, and smoking history) correlated with the intake of several food groups. Only intake of seabird was significantly correlated with BMI. Physical/sport activities correlated positively with shellfish, berries, vegetables, fruit and sweets and snacks but negatively with carbohydrate food and sauce. Furthermore, alcohol intake correlated positively with the intake of terrestrial animals and sweets and snacks. Current smoking was negatively correlated with intake of Greenlandic fish, shellfish, terrestrial animals, vegetables, and fruit, but positively with fast food. Cigarette smoking is a significant source of oxidative stress. One potential mechanism for its problematic health effects is that smoking contributes to endogenous oxidant formation through the inflammatory-immune response [43, 44]. Thus, it is worrying that current smoking is negatively associated with vegetable and fruit intake, as antioxidants consumed in the diet may act to mitigate against some of the adverse effects of cigarette smoking [45]. Other studies have shown that smokers have lower blood levels of antioxidants (ascorbic acid, α-carotene, β-carotene, vitamin A and E) than non-smokers [45], which might be due to both the oxidants in cigarette smoke and lower intake of antioxidants in smokers than non-smokers [45]. Smoking continues to be a public health problem in Greenland with more than half of the general population being current smokers [17] and 29–46% of the pregnant women in the ACCEPT cohort (2010–2015) smoked during pregnancy [18, 19]. A connection of smoking with an unhealthy diet may worsen the health problems further.

Overall, the food intake frequency of both the traditional and imported food groups in the present follow-up study was similar to previously reported frequencies in the ACCEPT birth cohort [18, 19].

However, the individual intake frequency of fruit has decreased with a median of 1.5 times/months from pregnancy to this follow-up study, and meat intake increased a median of 2.5 times/month. The decreased frequency in fruit intake may be due to several factors. It may be explained by a decreased fruit intake in the general population seen in the same period in the Population Health Surveys [16, 17], or the women may have been more aware of having a healthy lifestyle during their pregnancy. A Canadian study investigated dietary changes during pregnancy and found that pregnant women reported increasing intake of milk products, fruit, and sweet items and decreased or eliminated intake of caffeine, alcohol, and meats during pregnancy [46]. Increased fruit intake during pregnancy has also been reported in several other studies (reviewed by SE Hillier and EK Olander [47]). In line with the increase in meat intake in the present follow-up study from inclusion at pregnancy, the Canadian study found decreased or eliminated meat intake during pregnancy, which were mainly due to concern for the baby’s health, aversion, and nausea [46].

It is, to our knowledge, the first time that longitudinal dietary changes are reported in Greenland. The results may suggest that the women changed their diet during pregnancy and that the dietary habits have “normalized” to non-pregnancy habits after pregnancy. However, we do not have any measurements of the women’s intake before pregnancy, and it is possible that the women did not change diet during pregnancy but have changed the diet afterward.

To estimate the food intake of the different food groups, participants were asked to fill out the FFQ with 65 traditional and imported food items based on their intake during the last 12 months. FFQ is a widely used dietary assessment method, especially in epidemiological studies, and compared to other dietary assessment methods, it is relatively simple, cost-effective, and time-efficient. The self-reported food intake may be subject to recall bias, as the participants must remember their intake 12 months back in time, which can be difficult. The portion size was not included in the questionnaire, and it is possible that the individual amount consumed differs, even though we do not see any difference in the frequency. However, other studies have reported that the between-person variation in portion size was smaller and less important than the variation in frequency [48]. A previous study in Greenland found good agreement between a very similar FFQ and duplicate portions, even though there was a slight overestimation of traditional food and underestimation of sweets in the FFQ [4]. Due to lack of nutritional data for some food items and missing portion sizes and seasonal information, we have not been able to calculate the nutritional intake for the participants and compare with Nutrition Recommendations.

For some of the Greenlandic food items, intake may depend on hunting season and availability; thus, we tried to include the seasonal variation in the FFQ. However, season information was missing for 5–56% in the Greenlandic food items, and we were unable to include the information in the analyses. However, looking at the available seasonal information (Additional file 2: Table S2B), the season variation was limited for most items, but for some items, the differences between the percentage of eating the food item in spring/summer and autumn/winter were more than 15%. For seabirds (guillemot, common eider, and kittiwake), hare and grouse more participants reported intake in the autumn/winter than in spring/summer, while some marine mammals (minke whale, fin whale, porpoise/grind and hooded seal), champs, and some fish (capelin and trout) were eaten by more participants in the spring/summer than in the autumn/winter.

Due to limited funding and the logistic difficulties connected with data collection in the remote settlements in Greenland, we could only follow up on the families in the three biggest towns (Nuuk, Sisimiut, and Ilulissat). However, Greenlandic statistic reports show that 50% of the Greenlandic population live in these three towns, with a distribution (64% living in Nuuk, 19% in Sisimiut, and 16% in Ilulissat) comparable with the distribution in the present study (67% in Nuuk, 21% in Sisimiut, and 11% in Ilulissat) [49]. Thus, even though the present study results may not reflect the food intake for the total Greenlandic population, the study population represents at least half of the Greenlandic population with a distribution reflecting the population among the included towns.

In the general Greenlandic population, a more western lifestyle and higher socioeconomic status are generally seen in the towns, especially in Nuuk [17]. Due to the fact that only ACCEPT women in Ilulissat, Sisimiut and Nuuk were followed up in this study, they differed slightly from the ACCEPT birth cohort women (2013–2015) not included in this ACCEPT follow-up study; the women included in this follow-up study were older at delivery (median: 29.2 versus 27.4, p = 0.001), less likely to smoke during pregnancy (20.6% versus 38.6%, p = 0.001), had higher educational level (66.3% with technical college or university degree versus 42.3%, p < 0.001) and had higher marine food intake as indicated by the n-3/n-6 fatty acids ratio (median: 0.24 versus 0.21, p = 0.001) (data not shown). The women included in this study were similar to the non-included ACCEPT women with regard to BMI, parity, and alcohol consumption (before and during pregnancy) at inclusion (data not shown).