Obesity is a worldwide health problem associated with an increased risk of cardiovascular disease, type 2 diabetes mellitus, chronic kidney disease, several types of cancer, musculoskeletal disorders, and body image issues.1 Since the 1980s, there has been an overall increase in the prevalence of high body mass index values (BMI = weight in kilograms divided by height in meters squared). The increase among children and adolescents is of particular concern.2 It has been estimated that 41% of deaths and 34% of disability-adjusted life years related to BMI are due to cardiovascular disease in obese individuals. Furthermore, it has been shown that obesity in childhood increases associated cardiovascular risk factors (CVRFs) in adults, and thus there is a need for adequate surveillance systems to allow early diagnosis, monitor trends, and assess the efficacy of preventive interventions.3,4

BMI has long been accepted as an indicator for estimating body weight classification for screening, and epidemiological analysis, although it has limitations and has led to some controversy.5 In the infant and juvenile population, several age- and sex-specific cutoff points have been proposed according to reference growth standards.6 In older adults, BMI values of less than 23 have been associated with an increased risk of mortality.7

Obesity has a negative effect on CVRFs, hypertension, dyslipidemia, insulin resistance, and type 2 diabetes mellitus.1,2,8 Regardless of body fat, excess visceral abdominal adipose tissue is a reliable predictor of acquiring CVRFs.9

Obesity, abdominal obesity (AO), dyslipidemia, diabetes, and hypertension are recognized as biological risk factors for cardiovascular disease. These factors often cluster and lead to increased risk.9 Associations have been found between inadequate diet, a lack of physical activity, and sedentary lifestyles and an increased risk of obesity, cardiovascular disease, and the premature onset of other chronic conditions.10

This aim of this study was to provide a general estimate of the prevalence of overweight and AO in a random sample of the Spanish population older than 3 years and to analyze the influence of sociodemographic and lifestyle factors and their association with CVRFs. Previous studies have described the distribution of anthropometric indicators and prevalence by age bracket and sex.11,12

This analysis formed part of the ENPE study (Estudio Nutricional de la Población Española).11,12 The ENPE study was an observational cross-sectional study conducted in a representative population-based probabilistic sample (scope: persons more then 3 years of age residing in primary residences in Spain), which was selected using a random multistep procedure, with stratification of the units from the first step (census sections) according to autonomous community (k = 17) with 400 individuals per stratum (n = 6800). Second level: primary residences in each census unit. Residences were selected by random route sampling, 1 individual per household, with quotas controlled by age bracket (3-8, 9-18, 19-64, and ≥ 65 years), sex, and municipality size according to the population census (Spanish National Institute of Statistics, January 2014). Sampling error, 1.62%. The sample was weighted according to the Spanish population distribution by age bracket, sex, and municipality size. Variance was estimated by applying bootstrap techniques without correction for finite populations.

Information was collected by personal interview at the participant's home (May 2014-May 2015). The sample and the fieldwork logistics were designed and coordinated by a population studies company (SIGMA DOS).

The exclusion criteria for the anthropometric measurements were as follows: being pregnant, not being able to stand upright, or having limbs immobilized with nonremovable plaster casts. In total, 6800 individuals were included, of whom 3514 were women (51.7%).

The overall prevalence of obesity was 22.0% (95%CI, 21.0%-23.0%) (table 1) and that of overweight was 36.1% (95%CI, 34.9%-37.3%), which was significantly higher in males (41.8%; 95%CI, 40.1%-43.5%). Overweight and obesity increased with age after the age of 19 years, with higher prevalence in the age bracket of 65 years and older. In total 15.9% (95%CI, 12.6%-19.6%) of individuals in the 3 to 8 years age bracket were obese and 11.6% (95%CI, 9.3%-14.2%) of individuals in the 9 to 18 years age bracket were obese.

The estimated prevalence of AO (64.7%; 95%CI, 63.5%-65.8%) was higher in males (68.3%; 95%CI, 66.7%-69.9%), increased with age from the age of 19 years, and reached the highest rate in the age bracket of 65 years and older (91.6%; 95%CI, 90.0%-93.1%).

Factors associated with overweight and abdominal obesitySociodemographic factors

Table 1 shows the prevalence of each BMI category according to socioeconomic level, local population size, and autonomous community (adjusted for age). Table 2 shows the odds ratios (OR) of the prevalence of obesity and AO and their corresponding 95%CIs according to sociodemographic factors. A significant association was found between the prevalence of obesity and socioeconomic status, local population size, and geographic region. There was a lower risk of obesity in the 9 to 34 year age bracket than in the 3 to 8 year age bracket, but there was a significant increase in its risk after the age of 55 years. An inverse association was found between socioeconomic level and a lower risk of obesity in the North-Eastern, Central, and Eastern regions compared with the South.

Table 2.

Sociodemographic factors associated with the prevalence of obesity and abdominal obesity

	Obesity		Abdominal obesity
	OR (95%CI)	P	OR (95%CI)	P
Female	1		1
Male	1.13 (1.00-1.28)	.057	1.71 (1.52-1.93)	< .001
Age bracket		< .001		< .001
3-8 y	1		1
9-18 y	0.53 (0.36-0.78)	.001	0.60 (0.44-0.81)	.001
19-24 y	0.42 (0.27-0.64)	< .001	1.04 (0.77-1.40)	.814
25-34 y	0.62 (0.43-0.90)	.011	1.98 (1.50-2.63)	< .001
35-44 y	1.11 (0.80-1.55)	.518	3.80 (2.89-4.99)	< .001
45-54 y	1.14 (0.83-1.58)	.419	5.82 (4.42-7.67)	< .001
55-64 y	2.22 (1.61-3.08)	< .001	15.08 (10.94-20.79)	< .001
65-74 y	2.41 (1.73-3.34)	< .001	19.04 (13.55-26.76)	< .001
≥ 75 y	2.32 (1.63-3.28)	< .001	17.99 (12.19-26.53)	< .001
Socioeconomic level		< .001		< .001
High-medium high	1		1
Medium-medium	1.68 (1.30-2.17)	< .001	1.56 (1.28-1.89)	< .001
Low-medium	1.97 (1.55-2.49)	< .001	1.64 (1.36-1.96)	< .001
Low	2.13 (1.66-2.74)	< .001	1.76 (1.43-2.16)	< .001
Population size		.038		< .001
< 5000 pop	1		1
5000-15 000 pop	0.83 (0.67-1.03)	.090	0.81 (0.64-1.02)	.072
15 000-50 000 pop	0.85 (0.69-1.05)	.137	0.52 (0.41-0.64)	.000
50 000-200 000 pop	0.82 (0.67-1.01)	.058	1.11 (0.92-1.35)	.283
> 200 000 pop	1.04 (0.86-1.27)	.670	0.90 (0.74-1.10)	.314
Region		< .001		< .001
South	1		1
North-West	1.03 (0.82-1.29)	.801	0.64 (0.52-0.78)	< .001
North-East	0.63 (0.49-0.81)	< .001	0.66 (0.54-0.81)	< .001
Central	0.81 (0.66-0.99)	.037	0.65 (0.53-0.80)	< .001
East Central	0.83 (0.67-1.02)	.080	0.67 (0.55-0.82)	< .001
East	0.61 (0.49-0.76)	< .001	0.72 (0.59-0.87)	.001

95%CI, 95% confidence interval; OR, odds ratio of prevalence; pop, population.

An association was found between AO and sex, age, socioeconomic level, local population size, and geographic region (table 1). The risk of AO was higher in males than in females and progressively increased with age, especially after 55 years of age. An inverse association was found between AO and socioeconomic level.

Lifestyle

Four dietary patterns were identified: “meat-protein-pasta” (MPP), “sweets-cakes-dairy foods” (SCM), “Mediterranean diet” (MD) with more fish, legumes, yogurt, olive oil, fruit, and vegetables, and “salty snacks-dairy desserts-sweetened beverages” (SSDDSB). In persons older than 18 years, 4 lifestyle profiles were identified, which were mainly defined by the level of physical activity and sedentary time: a) not active-moderately sedentary/ MD+pattern, moving away from he SSDDSB pattern; b) very active-moderately sedentary/higher scores on the MD++, SCM+, and SSDDSB+patterns; c) not active-very sedentary/neutral diet; and d) not active-very sedentary/higher MPP+pattern score and lower SCM-pattern score.

The prevalence of obesity was significantly lower (P < .0001) in the active profile (17.7%) than in the other 3 profiles (24.5%, 22.7%, and 22.2%). The prevalence of AO was also significantly lower (P < .0001) in the active profile (62.7%) than in the other 3 profiles (72.0%, 69.8%, and 72.6%). Figure 1 shows the ORs of the prevalence of obesity and AO in models adjusted for age, sex, and socioeconomic status for the lifestyle and smoking profiles of persons older than 18 years. We estimated the ORs for obesity (OR = 0.82; 95%CI, 0.66-1.03) (figure 1A) and AO (OR = 0.85; 95%CI, 0.69-1.04) (figure 1B) of the very active-moderately sedentary profile and the not active-moderately sedentary profile. The ORs for obesity (OR = 0.72; 95%CI, 0.61-0.87) and AO (OR = 0.83; 95%CI, 0.71-0.97) were significantly lower in smokers than in nonsmokers.

Figure 1.

Odds ratios of the prevalence of obesity (A) and abdominal obesity (B) according to lifestyle profile and smoking. Odds ratios of prevalence adjusted for sociodemographic factors in persons older than 18 years. 95%CI, 95% confidence interval; MD, Mediterranean diet; OR, odds ratio.

(0.26MB).

Obesity, abdominal obesity, and cardiovascular risk factors

Figure 2A-C shows the ORs of the prevalence of obesity and AO in relation to the self-reported presence of dyslipidemia (hypercholesterolemia), hypertension, and diabetes in persons 18 years and older. In models adjusted for age, sex, socioeconomic status, smoking, and lifestyle profile, a postive association was found between overweight, obesity and AO and dyslipidemia (figure 2A) and hypertension (figure 2B). The risk of hypertension was lower in persons with a very active-low sedentary profile than persons with a not active-moderately sedentary profile. An association was found between obesity and AO and a higher risk of diabetes mellitus (figure 2C). The risk was lower in persons in the very active-moderately sedentary profile.

Figure 2.

Odds ratios of the prevalence of hypercholesterolemia (A), hypertension (B), and diabetes mellitus (C) according to lifestyle profile, smoking habit, weight classification, and abdominal obesity. Odds ratios of prevalence adjusted for age, sex, socioeconomic level, smoking habit, and lifestyle profile in persons older than 18 years. 95%CI, 95% confidence interval; MD, Mediterranean diet; OR, odds ratio.

(0.46MB).

The clustering of self-reported dyslipidemia, hypertension, and diabetes and obesity or AO was significantly higher in women and in persons of 65 years of age and older (table 3). An inverse association was found between the OR for the prevalence of CVRFs and socioeconomic level; in the North-East region vs the South region, OR = 0.32 (95%CI, 0.20-0.53); in smokers vs nonsmokers, OR = 0.60 (95%CI, 0.44-0.80); and in exsmokers, OR = 2.06 (95%CI, 1.48-2.86). The risk of clustering of CVRFs was lower in the very active-moderately sedentary profile/ MD++pattern than in the nonactive-moderately sedentary profile/MD+pattern.

Table 3.

Prevalence of the clustering of 3 or more biological cardiovascular risk factors and the odds ratio of prevalence according to sociodemographic factors and lifestyles in the population older than 18 years

	n	Prevalence	Odds ratio	P
		% (95%CI)	OR (95%CI)
Sex				.035
Female	342	12.6 (11.4-13.9)	1
Male	243	9.6 (8.5-10.7)	0.89 (0.72-1.11)
Age bracket				< .001
19-64 y	210	5.1 (4.5-5.8)	1
≥ 65 y	375	32.7 (30.0-35.4)	6.76 (5.38-8.48)
Socioeconomic level				< .001
High-medium high	37	5.4 (3.9-7.4)	0.44 (0.29-0.68)
Medium-medium	64	5.7 (4.4-7.1)	0.48 (0.34-0.68)
Low-medium	258	11.9 (10.6-13.4)	0.87 (0.68-1.11)
Low	222	19.3 (17.1-21.7)	1
Region				< .001
South	143	10.9 (9.3-12.7)	1
North-West	66	15.1 (11.9-18.6)	0.98 (0.67-1.43)
North-East	20	6.2 (3.8-9.0)	0.32 (0.20-0.53)
Central	155	11.3 (9.7-13.1)	1.01 (0.72-1.42)
Central-East	33	13.1 (9.2-17.5)	0.89 (0.62-1.28)
East	167	10.8 (9.3-12.4)	0.91 (0.63-1.32)
Population size				.094
< 5000 pop	97	11.5 (9.5-13.9)	1
5000-15 000 pop	119	12.2 (10.3-14.4)	0.87 (0.61-1.25)
15 000-50 000 pop	100	10.1 (8.4-12.1)	0.93 (0.65-1.33)
50 000-200 000 pop	113	10.2 (8.5-12.1)	0.80 (0.57-1.14)
> 200 000 pop	156	11.6 (10.0-13.4)	1.26 (0.91-1.75)
Smoking habit				< .001
No	403	13.1 (11.9-14.3)	1
Exsmoker	78	22.3 (18.2-26.9)	2.06 (1.48-2.86)
Yes	90	5.4 (4.4-6.5)	0.60 (0.44-0.80)
Lifestyle profile				< .001
1 Non-active-moderate sedentary/Mediterranean diet +/sweetened beverages -	133	16.0 (13.6-18.6)	1
2 Very active-moderately sedentary/Mediterranean diet++/ dairy foods-cakes+/sweetened drinks +	88	5.8 (4.7-7.0)	0.61 (0.44-0.84)
3 Not active-very sedentary/neutral diet	180	13.3 (11.6-15.2)	0.99 (0.74-1.32)
4 Not active-highly sedentary/protein-pasta++/dairy foods-cakes -	112	12.8 (10.6-15.1)	0.79 (0.58-1.08)

95%CI, 95% confidence interval; OR, odds ratio for prevalence; pop, population.

In the ENPE study sample, the estimated prevalence of overweight was 58.1% (95%CI, 56.9%-59.3%) and that of obesity was 22.0% (95%CI, 21.0%-23.0%). Overweight was higher in males and overweight and obesity increased with age after 19 years of age. The estimated prevalence of AO was 64.7% (95%CI, 63.5%-65.8%), higher in males, and increased with age after 19 years of age. These figures are similar to those reported in Ontario and Quebec (Canada) (6-79 years),17 and in Portugal (3 months-84 years).18 A study conducted in the United States (2015-2016) reported obesity prevalences of 39.6% in persons of at least 20 years.19 In the geographical area of the Organization for Economic Cooperation and Development (OECD), the estimated average of overweight was 60% and that of obesity was 25% (18.7% of persons younger than 20 years were overweight and 9.9% were obese), with obesity figures fluctuating between 3.6% (India) and 39.6% (United States).20

Recent analyses have projected a 16% increase in excess weight in Spain by 2030, with a 58% increase in direct health care costs.21

The Spanish National Health Surveys provide relevant public health information on weight, height, and health problems reported by participants. These data have been used to analyze the prevalence of obesity and associated factors in Spain22 and are of undoubted interest. However, these analyses underestimate the prevalence of obesity, with an error that has been calculated to be between 2.0 and 2.7 percentage points in men and 4.7 and 5.9 percentage points in women23 vs estimates based on measurements. The magnitude of obesity and its health, economic, and social impact necessitate a measurement-based quality surveillance system, which also includes quality information on related factors. Failing this, valuable information would be provided by well-designed population-based cross-sectional studies using well-designed protocols and individual anthropometric measurements.

The results of the present study are in line with those of previous studies, which found unequal distributions of obesity: its prevalence was higher in persons in lower socioeconomic brackets.22,24,25 Likewise, previous studies have found socioeconomic inequalities in the distribution of risk behaviors, such as poor dietary habits, low levels of physical activity, and tobacco consumption.24,25 Other studies have also found differences in the distribution of eating habits and physical activity according to autonomous community, which was mediated by the unequal distribution of educational and socioeconomic levels.26,27

A significant association was observed between the risk of obesity and AO and the size of the locality of residence. Other studies have found higher prevalences of obesity and AO in rural areas28 and have suggested social selection mechanisms as an explanation. It is important to pay attention to the urban environment in smaller localities and especially to the resources and health promotion initiatives available to reinforce nutritional literacy strategies and promote exercise together with initiatives that favor a less sedentary lifestyle in all age brackets.

Eating habits, physical activity, sedentary lifestyle, and hours of sleep have been associated with obesity.10,29 Some studies have analyzed the tendency of lifestyles to cluster30,31 and their relationship with obesity. The ENPE study identified 4 dietary patterns and 4 lifestyle profiles. Similar lifestyle patterns have been identified in studies on adults27,30 and children and adolescents.31 This type of analysis, which analyzes both lifestyle and diatetary pattern,is of particular interest because it can be used to design intervention strategies that are adjusted to the characteristics of the group under study. Risk behaviors tend to cluster, whereas higher levels of physical activity tend to cluster with better dietary profiles, although they may differ according to sociodemographic characteristics.30,31 The present study found an association between the more active-moderately sedentary profile/MD++pattern and a lower risk of obesity, dyslipidemia, hypertension, and diabetes.

Spanish studies have used objective indicators to analyze the clustering of CVRFs (obesity, AO, dyslipidemia, hypertension, diabetes).27,32,33 Our study found that there was a higher risk of clusters of 3 or more CVRFs in exsmokers than in smokers and persons with a more active-moderately sedentary lifestyle. Other studies have also found a higher prevalence of CVRF clusters in nonsmokers and former smokers.32,34 Being at increased cardiovascular risk could favor smoking cessation in smokers (exsmokers) after receiving medical advice (reverse causality). It is also known that smoking cessation is associated with weight gain, risk of obesity,34 and diabetes.35 Patients with concurrent obesity, dyslipidemia, hypertension, and a smoking habit could have refused to participate or even died and thus would not have been included in the prevalent cases.

Positive associations have been found between smoking and obesity, and even between obesity and the number of cigarettes consumed, especially in adolescents and young adults.34 It would be desirable to reinforce smoking prevention strategies and promote healthy lifestyles in adolescents and young adults. Furthermore, smoking cessation interventions should also include measures aimed at adjusting energy intake and promoting the practice of physical activity to avoid or limit weight gain.

In the Spanish population, the prevalence of obesity and AO is high, the prevalence of AO is higher in men, both conditions increase with age, and both are inversely associated with socioeconomic level. Lifestyle profiles combining a higher level of physical activity, a moderately sedentary lifestyle, and a Mediterranean dietary pattern are associated with a lower risk of obesity, AO, and CVRFs. Early intervention strategies that promote better eating habits, increased physical activity, and reducing sedentariness are key to the prevention of these problems. The prescription of dietary advice and physical exercise is also essential in the treatment of affected individuals.

The ENPE study (Estudio Nutricional de la Población Española) fieldwork was funded by the Eroski Foundation through an agreement with SPRIM-España and the Spanish Society of Community Nutrition (SENC). The sponsor was not involved in the study design, data collection, analysis, interpretation of the results, writing the manuscript, or the decision to publish the results.

None declared.