Background
As obesity has taken epidemic proportions, higher prevalence of obesity is observed also in pregnant women [
1-
3]. Maternal overweight or obesity during pregnancy can lead to several adverse outcomes for both mother and foetus. Obese women are more likely to have gestational diabetes mellitus, hypertension, pre-eclampsia, other obstetric complications and caesarean section [
1,
4,
5]. Pre-pregnancy obesity is further associated with increased risk of adverse pregnancy outcomes such as stillbirth [
6], preterm delivery [
3], birth injuries [
7], and that the new-borns become large for gestational age [
1,
4,
5] with subsequent increased risk for childhood obesity [
1]. Maternal gestational weight gain (GWG) is also related to these adverse pregnancy and birth outcomes [
8,
9], including an increased risk of macrosomia at birth [
9], and increased risk of childhood obesity [
10,
11]. Thus, obesity in early pregnancy and also excessive GWG are known to have consequences for pregnancy and birth outcomes and the risk of childhood obesity in the next generation.
It is well known that individuals with lower education and socioeconomic position are more often overweight or obese in comparison to people with higher education or socioeconomic position. This social inequality is also observed in pregnant women [
5,
12,
13]. A steepened social gradient has been implied in the Nordic countries but the evidence for increased social inequality of overweight and obesity is inconsistent [
14]. Thus, further knowledge about time trends of socioeconomic inequalities in overweight or obesity is needed.
An urban–rural gradient of obesity has also been observed [
12,
15,
16], with lower risk of obesity in urbanized areas. Among Swedish pregnant women, living in rural areas has been shown to almost double the odds of obesity after taking occupation and education into account [
12]. Also in young Swedish males an urban–rural gradient is observed taking socioeconomic position, intelligence quotient and parental education into account [
16]. The risk of obesity is further related to country of birth and data from the Swedish Medical Birth Register indicate that women with non-Nordic origin have higher prevalence of overweight [
3], but more specific results on country of birth are lacking. Other European studies indicate higher pre-pregnancy BMI among women from African and Middle Eastern countries compared to the majority population in each country [
17].
The main aim was to explore time trends of weight status among pregnant women in Sweden and to investigate if the educational gap in overweight or obesity has increased, decreased or remained stable during the last two decades. Additional aims were to explore differences in overweight and obesity according to area of residence and country of birth.
Discussion
The prevalence of overweight and obesity among Swedish pregnant women has increased since 1992, especially among those with low education. Women with low education or living in sparsely populated areas are at higher risk of obesity in early pregnancy.
In 2008–2010, 25% of Swedish pregnant women were overweight and 12% were obese at their first visit to the antenatal care clinic. This is slightly higher than pre-pregnancy weights observed in other countries [
24-
27]. Discrepancies are that most previous studies investigated somewhat older data, and that weight was assessed early in the first trimester instead of pre-pregnancy weight in the present study. A GWG rate of 0.22 kg per week has previously been estimated during the first trimester [
28] and the most frequent time point for the initial visit to the antenatal clinic is at ten weeks of gestation. If a 2.2 kg weight gain before the first antenatal care visit was anticipated for all participants, estimated pre-pregnancy obesity prevalence among nulliparous women was 1.0-1.6 percent units lower than the prevalence observed in Table
2. According to the Danish Medical Birth Register, 21% of the pregnant women had overweight and 12% had obesity during 2004–2010 [
24]. In the Norwegian Mother and Child Cohort Study conducted between 1999 and 2008, 22% of the women becoming pregnant were overweight and 9% were obese [
25].
Our results showed a slower increase in prevalence of obesity during the 21
th century, which is in accordance with previous findings in Swedish populations [
29-
31]. However, Swedish data on whether the socioeconomic gradient of obesity is changing are inconsistent [
14]. A previous Swedish study has indicated a narrowing of the socioeconomic gap for adults aged 25–44 years between 1988–1989 and 1996–1997, whereas it seemed stable for adults 45–64 years [
32]. In contrast, a higher increase in prevalence of obesity in lower social classes has been suggested in Stockholm County during the 21
th century [
30]. Increased socioeconomic gap in overweight and obesity has also been shown among young Swedish men between 1970 to 2000 [
33]. An increased socioeconomic difference in obesity prevalence from 1990–1995 to 2002–2007 was further observed in the North of Sweden [
29]. Using education as a measure of socioeconomic position, the present study indicates that the socioeconomic gap in obesity is increasing, both in relative and absolute terms. In contrast, the socioeconomic difference in overweight seems to be stable over time. We can only speculate in why the socioeconomic gap of obesity appears to increase but the increasing physical activity in Sweden seems to be mainly among individuals with higher education [
29]. Further, individuals with higher education may be moving to urbanized areas, with its beneficial contextual factors, to a higher degree [
29]. It may also be speculated that individuals without university education are more excluded from the labour market today or that health literacy has become more central in the health care system.
In line with other Swedish studies [
12,
15,
16,
29], an urban–rural gradient of overweight and obesity was observed among pregnant women independently of educational level. The sample size in the present study allowed for a more detailed categorisation of residential area than the often used urban/rural classification. Interestingly, the risk for overweight or obesity seems to increase linearly with decreasing population density of municipality of residence based on comparisons of risk ratios. Possible explanations may be that habitants in rural areas are more dependent on motorized transportation and higher accessibility to training facilities in urbanized areas. Also other contextual factors could contribute, such as that a larger part of the population is highly educated in urbanized areas, irrespective of the own educational level.
Our results show both increased and decreased risks of overweight or obesity in different groups of non-Swedish born pregnant women compared to Swedish born pregnant women. Women born in Africa, the Middle East, and Latin America had higher risk compared to native Swedes, whereas women born in other countries had lower risk. These risks remained after adjustment for education. Other studies [
27,
34,
35] confirm ethnical differences in obesity prevalence, especially among immigrants from Middle East or Latin America. Compared to native Swedes, more than twice as high obesity prevalence has been observed among women from Middle East [
35], Turkey [
34] and Chile [
34]. Likewise, in a Dutch study [
27], ethnical differences in GWG and postpartum weight retention were observed, with Turkish women being at higher risk for retaining weight. Women born in other Nordic countries than Sweden had a slight over risk of obesity. Previous studies have shown higher BMI and body fat percentage among Finnish-born women compared to native Swedes [
36,
37]. Possible factors explaining the differences in obesity between countries of birth could be socioeconomic inequalities, difficulties to be included on the labour market, diverse lifestyle, genetic susceptibility, language barriers and cultural attitudes to sickness and health care [
35].
Large proportions (57-63%) of the overweight and obese women gained more in weight during pregnancy than recommended by the IOM [
8,
20] and were at increased risk of short-term and long-term complications to the mother and the offspring during pregnancy, delivery and the post-term period. Our results indicate that further research is needed on evaluations of primary prevention interventions which might be implemented in routine antenatal care to limit excessive GWG among Swedish pregnant women. While the educational differentials in body weight in early pregnancy were strongly inversely related to maternal level of education, GWG showed less strong associations. In line with ours, a recent study showed that excessive GWG was more prevalent among low educated normal weighted women whereas education did not seem to protect against excessive GWG among overweight and obese women [
13]. In fact, we observed a slightly higher risk of excessive gestational weight gain among higher educated overweight or obese women, also after adjusting for pre-pregnancy BMI.
The present study has strengths and limitations. It is a strength that it is nationwide by including almost all pregnancies progressing to labour in Sweden during the investigated time period. Due to the large number of women, we had sufficient power to investigate residential area and country of birth at more detailed levels. Since nearly 75% of the Swedish women have at least one child [
22], the results are believed to be generalizable to the fertile Swedish female population. The study is limited by that the Swedish Medical Birth Register does not contain data on weight before conception. Weight was assessed at the first visit to the antenatal care clinic. Most of these visits are scheduled between gestational week 8 to 12 [
18].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors have made substantial contributions to conception, design and interpretation of data. SL performed the statistical analyses. HB wrote the manuscript. FR provided critical feedback on the manuscript. All authors read and approved the final manuscript.