Obesity continues to increase in the majority of the population in mid-Sweden—a 12-year follow-up

Abstract

Background: The aim was to investigate trends in the prevalence of obesity by age and level of education in the general population in mid-Sweden from year 2000 to 2012. Methods : A postal questionnaire was sent to a random population sample aged 25–74 years in years 2000, 2004, 2008 and 2012. The overall response rates were 67%, 65%, 60% and 53%, respectively, and the study included 29 017, 27 385, 25 910 and 24 152 respondents, respectively. Obesity (BMI ≥ 30 kg/m ² ) was based on self-reported weight and height. Results: The age-standardized prevalence of obesity increased from 13% to 17% in women and from 12% to 17% in men between 2000 and 2012. Obesity increased in all age groups from 2000 to 2008 and continued to increase among the middle aged (45–64 years) between 2008 and 2012. The socioeconomic gradient in obesity changed during the study period since the absolute increase in obesity was steepest at the middle educational level. In 2012, the prevalence of obesity was almost twice as high at both middle and low educational levels compared with high educational level. The ‘true’ prevalence of adult obesity, corrected for self-reported weight and height, was around 20% in 2012 for both men and women. Conclusion: In the majority, among the middle-aged and those with secondary education, the prevalence of obesity continued to increase even between 2008 and 2012.

Introduction

The prevalence of overweight and obesity has increased dramatically worldwide since the 80s and excess body weight and related comorbidities have become a major public health issue in both developed and developing countries. ^1,2 Obesity is a chronic disease that is very difficult to treat, and comprehensive preventive actions are necessary to tackle the global obesity epidemic. However, several studies suggest that the increase in obesity has levelled off since the early 2000s, especially in the Western countries. ^3,4 In a review, Rokholm et al.³ concluded that the obesity epidemic among children and adolescents has reached a plateau in several parts of the world. Among adults, the results are inconsistent with some studies showing stability, while others show continued increase. Mixed results are found both across countries and among population subgroups within countries, e.g. stability was less evident in groups with low socioeconomic status. ^3,5 Studies reporting a break in obesity prevalence increase have relatively short follow-up periods and analysis of long-term trends has shown that the obesity epidemic has developed progressively with periods of stability followed by new increases to higher levels. ^3,5

Reliable information on country-specific trends in overweight and obesity are needed to assess and monitor the effects on public health and to evaluate the results of public health promotion and prevention efforts. Also, it is important to identify population subgroups with the highest obesity prevalence, or the fastest increase, in order to target interventions to those who need it most. In Sweden, the prevalence of adult obesity doubled from around 5% to 10% during the 80s and 90s. ⁶ However, a national study found no increase between 2000/2001 and 2004/2005, although significant increases were observed in some specific subgroups. ⁷ Another national study showed that the prevalence was stable also during the period from 2004 to 2008. ⁸ The prevalence of obesity in Sweden year 2008 was estimated to 12% for men and 11% for women in the age group 16–84 years. However, these results were based on self-reported weight and height that are likely to underestimate obesity. ⁹

Studies of obesity prevalence in the middle-aged population in Northern Sweden between 1990 and 2007 also showed a plateau around year 2000 but only among those with high education in the most urbanized areas, while obesity continued to increase in other subgroups. ¹⁰ A stabilization of obesity rates among those with higher socioeconomic status has been reported from other European countries. ⁵ Regional studies in Stockholm County showed that the prevalence of obesity was stable during 2002–06 but increased between 2006 and 2010, ¹¹ indicating a possible upsurge to new higher levels after a period of stability.

Prevalence of obesity in the adult Swedish population increases with age and reaches a peak among men aged 55–64 and women aged 65–74 years. ^7,11 Obesity is also more common among those with lower education in Sweden ^7,10,12 as in most other high-income countries. ^13,14 The prevalence of obesity has increased in all subgroups of the Swedish population, but the social gradient with higher obesity prevalence in groups with lower socioeconomic status has remained fairly stable from the 80s to the mid-2000s. ¹⁵ Some studies, however, indicate a growing gap in obesity rates between those with high and low levels of education. ^10,16,17 The aim of this study was therefore to investigate trends in the prevalence of obesity by sex, age and level of education in the general population in mid-Sweden from year 2000 to 2012.

Methods

Study population

The study is based on a postal survey questionnaire sent to a random population sample in the years 2000, 2004, 2008 and 2012. The sampling frame was the population register at Statistics Sweden, the statistical administrative authority in Sweden, covering all inhabitants of the study area. Data collection was discontinued after two postal reminders. The overall response rates were 67%, 65%, 60% and 53%, respectively. The area investigated covers four counties (Sörmland, Uppsala, Västmanland and Örebro) including 39 municipalities with about 1 million inhabitants in the central part of Sweden. The age range was 25–74 years and the study population included 29 017, 27 385, 25 910 and 24 152 respondents, respectively.

Measures

Body mass index (BMI) was based on self-reported weight and height and categorized according to the WHO guidelines ¹⁸ to underweight (BMI < 18.5 kg/m ² ), normal weight (18.5≤ BMI < 25 kg/m ² ), overweight (25≤ BMI < 30 kg/m ² ) or obesity (BMI ≥ 30 kg/m ² ). Since some recent studies ^19–21 have found increasing trends especially at the upper end of the BMI distribution, we also calculated trends separately for those with 30 ≤ BMI < 35 kg/m ² and BMI ≥ 35 kg/m ² (severe obesity).

Sex, age, educational level and country of birth were based on official registers. Levels of education were categorized into compulsory school or equivalent education for 9 years or less, secondary education (10–12 years of education) and post-secondary education (>12 years of education). Country of birth was categorized into those born in Sweden, other Nordic countries, other European countries and outside Europe. Smoking habits were derived from the questionnaire and dichotomized into cigarette smoking daily and not daily.

Statistical methods

The development of obesity prevalence in the population from 2000 to 2012, including comparisons over educational levels, was estimated by age-standardized proportions with corresponding 95% confidence intervals. Analyses were stratified by sex. All analyses were conducted in IBM SPSS Statistics version 22.

Since the prevalences of obesity in this study are based on self-reported weight and height, which are likely to underestimate obesity, ^9,22,23 we also calculated corrected estimates for the overall prevalences for years 2000 and 2012. We used algorithms suggested by Nyholm et al.⁹ based on a Swedish adult population. They showed that age and measured BMI were important factors for misreporting height, weight and BMI in both men and women, whereas other variables such as education, physical activity and smoking did not improve the prediction of obesity. Therefore the algorithms take into account that the difference between BMI based on measured weight and height and BMI based on self-reported weight and height (BMI ^SR ) increases with increasing BMI and age. The corrected BMI (BMI ^C ) was calculated as follows:

For men: BMI ^C = −0.202 + 1.005 × BMI ^SR + 0.014 × age.

For women: BMI ^C = −0.713 + 1.023 × BMI ^SR + 0.019 × age.

Ethical approval

The respondents agreed that official data records would be linked to the data in the questionnaire by providing informed consent. Statistics Sweden added registry data and returned unidentified files to the counties. This was done under the jurisdiction of Swedish law of official statistics (2001:99 6§) and the law of secrecy (1980:100 9 chap. 4§), and the Declaration of Helsinki. The Ethical Review Act of Sweden (2003:460) did at the time of the data collection in years 2000–08 not require an approval of an ethics committee since the data are anonymous. In 2012 the survey was approved by The Regional Ethical Review Board of Uppsala (EPN 2012/256).

Results

Table 1 lists the background characteristics of the study population by year. In general, there was a shift towards older ages and higher education in the study population from year 2000 to year 2012. The distribution of country of birth was relatively stable. The prevalence of daily smoking decreased from 17% in 2000 to 12% in 2008 but was stable between 2008 and 2012.

Table 1

Characteristics of the study population by year

Year	2000	2004	2008	2012
N	29 017	27 385	25 910	24 152
Sex (%)
Women	53.0	53.7	54.1	54.3
Men	47.0	46.3	45.9	45.7
Age group (%)
25–34 years	15.9	15.7	13.5	11.2
35–44 years	18.7	17.8	17.4	15.5
45–54 years	22.1	19.7	18.6	19.1
55–64 years	21.7	23.0	24.3	23.3
65–74 years	21.5	23.8	26.3	30.9
Educational level (%)
Post-secondary education	12.9	25.7	27.0	36.5
Secondary education	58.8	50.7	51.2	46.1
Compulsory education	28.3	23.6	21.8	17.4
Country of birth (%)
Sweden	87.8	87.5	87.3	86.4
Other Nordic country	6.3	5.8	5.6	5.7
Other European country	3.0	3.0	3.7	3.6
Outside Europe	2.9	3.6	3.4	4.3
Daily smoker (%)
Yes	17.3	15.9	12.4	12.5
No	82.7	84.1	87.6	87.5

The age-standardized prevalence of obesity increased statistically significantly from 13% in women and 12% in men to 17% in both women and men between 2000 and 2012 ( table 2 ). The increase in obesity was, however, larger between 2000 and 2008 than between 2008 and 2012. The prevalence of severe obesity (BMI ≥ 35 kg/m ² ) increased as well between 2000 and 2012 but showed little increase between 2008 and 2012. In both women and men, there was a statistically significant decrease in the proportion of normal weight persons, whereas the prevalence of overweight (25 ≤ BMI < 30 kg/m ² ) did not change.

Table 2

Age-standardized distribution (%) of BMI among women and men aged 25–74 years by year

BMI (kg/m 2 )	2000	2004	2008	2012	Difference 2000–12
Women
<18.5	1.7	1.5	1.8	1.6	−0.1
18.5–24.9	54.6	54.4	50.0	50.8	−3.8 *
25–29.9	31.1	29.9	31.7	30.7	−0.4
≥30	12.7	14.1	16.5	16.9	4.2 *
30–34.9	9.6	10.5	11.8	12.2	2.6 *
≥35	3.1	3.7	4.8	4.7	1.6 *
Men
<18.5	0.4	0.3	1.0	0.4	−0.1
18.5–24.9	41.2	39.4	36.3	35.3	−5.9 *
25–29.9	46.5	46.6	46.5	47.5	1.0
≥30	11.8	13.7	16.3	16.9	5.1 *
30–34.9	9.8	11.1	12.7	13.2	3.4 *
≥35	2.0	2.6	3.5	3.7	1.7 *

*Statistically significant ( P < 0.05).

Obesity increased in all age groups from 2000 to 2008 but decreased somewhat, although not statistically significantly ( P > 0.05), between 2008 and 2012 in the youngest age group 25–34 years ( figure 1 ). This applies both to women and men. In the age groups 35–44 and 65–74 years, no increase was seen between 2008 and 2012. Among women and men aged 45–64 years, there was a statistically significant ( P < 0.05) increase also between 2008 and 2012.

Figure 1

Crude prevalence (%) of obesity among women and men by age group and year

Age-standardized prevalence of obesity increased between 2000 and 2012 at all educational levels but was more common among those with low education compared to subjects with high education in all four surveys. The increase was, however, steepest—about 8 percentages—at the middle educational level ( figure 2 ). In 2012, the prevalence of obesity was roughly equal among persons with low and middle level of education ( P > 0.05), whereas the prevalence was significantly lower among those with high education ( P < 0.05). This applies both to women and men. Between 2008 and 2012 there was no increase in obesity either in men with low or high education, whereas in women with high education, the increase was statistically significant ( P < 0.05).

Figure 2

Age-standardized prevalence (%) of obesity among women and men aged 25–74 years by level of education and year

The corrected age-standardized prevalence of obesity was 16.0% for women and 15.0% for men in year 2000 and the corresponding prevalence in year 2012 was 20.8% for both women and men (data not shown). Thus, the corrected prevalence was about 3% higher in year 2000 and about 4% higher in 2012 than uncorrected estimates based on self-reported weight and height.

Discussion

In our study of a general adult population in the central part of Sweden, the age-standardized prevalence of obesity increased between 2000 and 2012, while the prevalence of overweight was stable. Similar obesity prevalence and trends were seen in women and men. Obesity increased in all age groups from 2000 to 2008 and continued to increase between 2008 and 2012 among the middle-aged, while a levelling-off was observed in the youngest and oldest age groups. Obesity prevalence increased at all educational levels. The increase was, however, steepest at the middle educational level, and in 2012, there was no difference in the prevalence between the middle and the low educational levels.

Previous studies of obesity trends among adults in Sweden have given mixed results. Nationally representative studies showed no increase between 2000 and 2008, ^7,8 while regional studies with larger samples showed partly different trends. In Stockholm County, stable obesity prevalence was observed during the first part of the 2000s but rising prevalence in the latter part both among women and men. ¹¹ In Northern Sweden, obesity prevalence increased between 1990 and 2007 in most subgroups of the middle-aged population, while a plateau was observed in the early 2000s among highly educated people living in urbanized areas. ¹⁰ A register-based national Swedish study including more than 1.5 million pregnant women showed a steady increase in obesity prevalence (based on weight and height from the first visit to the antenatal care) during 1992–2010, but the increase tended to slow down during the latter part of the 2000s. ¹⁶ Thus, it is likely that the prevalence of obesity in Sweden has increased at different rates and patterns for different population groups and regions.

In our study, a continuous increase in the prevalence of obesity was seen in the total age group 25–74 years between 2000 and 2012; however, a somewhat slower pace was observed between 2008 and 2012 due to levelling-off in the youngest and oldest age groups. It is, however, too early to say whether the latter trend is stable and reliable or maybe a consequence of declining participation rates, especially among younger age groups in the 2012 study. In the USA, the increase in obesity among adults levelled off before 2010, ⁴ but the increase in morbid obesity (BMI ≥ 40 kg/m ² ) has continued. ²¹ Recent studies from Australia and Canada ^19,20 have also reported increases during the 2000s especially at the upper end of the BMI distribution; however, our results showed no increase in severe obesity between 2008 and 2012.

In contrast to the nationally representative studies, our study and several other studies indicate a steady increase in obesity prevalence during the 2000s in large subgroups of the population. The population of the study area in mid-Sweden covers many types of smaller and larger, rural and urbanized municipalities but does not include any metropolitan region such as Stockholm and is therefore not necessarily representative of the Swedish population as a whole. The prevalence of obesity is estimated to be about 2% higher in the study area, compared to the corresponding national prevalence. ²⁴

Since the prevalence estimates in this study are based on self-reported weight and height, which are likely to underestimate obesity, we carried out a sensitivity analysis using an algorithm based on a Swedish validation study. ⁹ The corrected estimates indicate that the ‘true’ prevalence of obesity is probably around 4% higher in 2012 than uncorrected estimates. The corrected obesity prevalence for 2012 was around 20% for both men and women, which is alarmingly high and in line with the prevalence estimates of 18.9% for Swedish men and 19.8% for women presented in the Global Burden of Disease Study for 2013. ¹ In addition, the increase in obesity prevalence could be even higher than shown here since the difference between measured and self-reported BMI increases with higher BMI. ⁹ Also, the response rate decreased during the study period from 67% to 53%, which may have caused a selection bias also leading to underestimation of obesity. ⁵ However, an earlier non-response study, carried out in a sub-area of this study area, showed that there was no significant difference in obesity prevalence between respondents and non-respondents. ²⁵ A tendency with decreasing response rates over time has been observed also in other population studies on obesity trends. ^7,8,11

The social gradient of obesity is well known, with higher socioeconomic groups having lower obesity prevalence in high-income countries. ^{13,14,26–28} Obesity trends within different socioeconomic groups have shown mixed results, both in Sweden ^10,16,17 and other European countries, ⁵ with some studies showing increasing socioeconomic gradients. A newly published review of studies in the Nordic Region concluded persisting social gradient in obesity in all Nordic countries. ¹⁵ Our results confirm the social gradient in 2000 when using educational level to measure socioeconomic status. An increasing trend in obesity was observed at all educational levels from 2000 to 2012. The absolute increase was, however, largest (8 percentages) at the middle educational level, which led to a change in the socioeconomic gradient. In 2000, there were distinct differences in obesity prevalence between the three levels of education. However, in 2012, the uncorrected prevalence of obesity was about 20% at both the low and middle educational levels, compared to 11–13% in the high education group. This finding emphasizes that obesity is not only a health problem among those with low education but also a major health concern among those with mid-level education, who represent about half of the total adult population.

The differences in obesity prevalence between socioeconomic groups may reflect differences in lifestyle habits such as dietary habits, physical activity, smoking and alcohol consumption. ²⁹ Differences in other factors such as weight control practises, ³⁰ reproductive history ³¹ and psychosocial factors ³² may also contribute to socioeconomic differences in obesity levels. Furthermore, the association is also known to be bidirectional since obese persons meet more difficulties in attaining high socioeconomic positions, e.g. due to discrimination in the working life. ^33,34

Women and men showed similar increasing trends of obesity between 2000 and 2008 at different educational levels. There were slightly different trends from 2008 to 2012, but these were in general not statistically significant.

A shift towards higher education in the study population was seen. This is partly due to a shift in the criteria for classification of education in year 2000 ³⁵ and partly due to a real shift towards higher education in the population between 2000 and 2012. ³⁶ The declining response rates in our study may also have had an impact since persons with low education tend to participate in population studies to a lesser degree than persons with higher education. With a shift towards higher education, a smaller increase in the overall obesity prevalence would have been expected. Also, a decrease in smoking prevalence occurred in the study population until 2008. Smoking cessation is associated with a moderate increase in body weight, ^37,38 but it is unlikely that the observed decrease in smoking prevalence would be a major explanation to the increasing trend in obesity in the study population. Previous studies have reported ethnical differences in obesity prevalence, especially among immigrants from Middle East or Latin America as well as Finnish-born individuals, compared to native Swedes. ^16,39,40 The proportion of persons born outside Sweden in our study did, however, not increase during the study period and can therefore not have affected the trend estimates.

A strength of our study is that it is based on a large general population and uses data from four consecutive, cross-sectional surveys carried out using the same methodology. The study covers a long time period from 2000 to 2012, while most previous studies reporting trends in obesity prevalence during the 2000s have had relatively short follow-up periods. Another strength is that we report trend estimates of severe obesity since recent population trends on severe obesity in Sweden have not been available. We also present obesity prevalence figures corrected for self-reports. These estimates are probably more reliable than the figures that have been reported in other Swedish population studies. Even though this study is not necessarily nationally representative, it is likely that similar overall trends could be seen in the Swedish adult population as a whole, i.e. the high prevalence of obesity and the alarming increase of obesity among persons with secondary education. To verify this conclusion, nationally representative studies are required.

Conclusion

Obesity increased in all age groups and at all educational levels from 2000 to 2008 and continued to increase in the majority, among the middle-aged and those with secondary education, even between 2008 and 2012, indicating that obesity is a growing burden to public health in this large general population in mid-Sweden. The socioeconomic gradient changed since the increase was steepest at the middle educational level. In 2012, the prevalence of obesity was almost twice as high at both middle and low educational levels compared with high educational level. Prevalence estimates corrected for self-reported weight and height indicate that the ‘true’ prevalence of obesity is around 3–4% higher than uncorrected estimates. In 2012, the ‘true’ prevalence of adult obesity was around 20% for both women and men.

Funding

This study was supported by the County Councils of Västmanland, Sörmland, Uppsala and Örebro.

Conflicts of interest : None declared.

Key points

• Previous studies of recent obesity trends among adults in Sweden, as well as in many other Western countries, have shown mixed results and have mainly had short follow-up periods.

• Our results indicate that obesity increased in all age groups from 2000 to 2008 and continued to increase in the majority, among the middle-aged and those with secondary education, even between 2008 and 2012 in mid-Sweden.

• The socioeconomic gradient changed since the absolute increase was steepest at the middle educational level. In 2012, the prevalence of obesity was almost twice as high at both middle and low educational levels compared with high educational level.