Updated Information On Prevalence Of Overweight

Published Date: 02 Nov 2017

Abstract

We aim to provide updated information on prevalence of overweight and obesity in Spanish adult women, and to also provide references for body composition. We considered data from a survey conducted annually between 2009 and 2010, on 4,013 subjects aged 18 years or more (mean age 53.42±11.99 years), representatives of the Spanish adult women, including measured information on height and weight. A subgroup of 2,224 women conducted a bioelectrical body impedance analysis of body composition. Segmental fat mass (Kg) and fat free mass (Kg) were measured with the Tanita Body Composition Analyzer (Model no. BC-418 MA) and percentile curves were constructed. The World Health Organization (WHO) criteria based on Body Mass Index (BMI) was used for the definition of overweight and obesity. Overweight was present in the 38.4% (95% CI 36.9-39.9%; n=1,539) of the sample, and the prevalence of obesity was of 24.6% (95% CI 23.3-26.0%; n=989). A total of 1,485 women (36.9%; 95% CI 35.4-38.4%) were considered normal according to the BMI. Prevalence of overweight and obesity were significantly increased with age. We have observed an elevated prevalence of overweight and obesity in Spanish adult women. Reference values for fat percentage, fat mass and fat free mass are provided.

Introduction

In the last several decades, the alarming increase in obesity has become one of the most important topics in human biology. Obesity is the outcome of complex interactions between genetic and environmental factors [1]. The latter are mostly associated with a modern lifestyle [2]. The acute rise in the rates of obesity [3–6] is common between men and women and rich and poor people from different ethnicities [7]. Therefore, obesity is a public health concern that is associated with an increased risk for developing hypertension, lipid disorders, type 2 diabetes, heart disease [8], stroke [9] and cancer [10]. Obesity has also been associated with an increased risk for early all-cause mortality [11].

Body mass index (BMI) [12] is widely used as an anthropometric estimate of general adiposity, but the failure to identify differences in body composition and body fat distribution has limited the usefulness of BMI. Therefore, the combination of BMI and body fat distribution might be more useful as a clinical prognostic indicator and a diagnostic criterion for obesity. A number of methods are used for measuring body composition and body fat distribution [13–15]. Among the gold-standards for body composition analysis are computed tomography (CT) and dual-energy X-ray absorptiometry (DXA). The bioelectrical body impedance analysis (BIA) is a simple, low-cost, safe, accurate and reproducible method for body composition analysis [16]. It has shown great potential in epidemiological and clinical studies [17,18]. BIA approach for the measurement of the body composition has additional advantages: It requires low-cost instrumentation and is easily applicable in practice. However, several factors might limit the valid application of BIA analysis as body geometry and body water distribution could interfere. These factors might affect on the validity of the method that result in either the underestimation or overestimation [19] of the fat mass.

To our knowledge, there is still limited information available on body composition changes determined by BIA and stratified by age in the female Spanish population. The aims of the present study were i) to assess the prevalence of the overweight and obese in the area of Extremadura, southwestern Spain; ii) to describe and to provide reference ranges for the age-related changes in body composition and iii) to correlate the body composition patterns with the BMI in a large cohort of Spanish women.

Methods

Ethics statement

This study was approved by the Research Ethics Committee of the University of Extremadura. All the participants gave written informed consent prior to the study.

Subjects

This study was a cross-sectional study and included a total of 4,013 Spanish women between 18 and 87 years of age. The inclusion criteria were that the subjects should be ambulatory and that, by self-reports, their body weight should not have changed by >2 kg over the previous 6 months. The exclusion criteria included severe organ failure and conditions and medications known to affect body weight and/or body composition. A sub-group of 2,224 women (BIA cohort) had a BIA to analyze body composition.

Anthropometry

The subject’s body weight was measured using a digital scale. Height was measured without shoes using a Harpenden stadiometer with mandible plane and parallel to the ground. We calculated the BMI using weight in kg divided by height in m2 (kg/m2).

Body composition measurements by bioelectrical impedance analysis

The BIA was performed in the early morning after an overnight fast for at least 12 h. The subjects were required to adhere to standard BIA testing guidelines described by the 2004 European Society for parenteral and Enteral Nutrition (ESPEN) [16]. The bioelectrical impedance was measured in subjects wearing light-colored clothes and standing erect with their bare feet on the analyzer's footpads. The feet were cleaned with soap and water and air-dried prior to the BIA procedure. A Tanita BC-418 MA Segmental Body Composition Analyzer (Tanita Corp., Tokyo, Japan) was used. Specific data for body composition calculations included age, sex and body type (athletic, normal). The analyzer measurements were made using 8 polar electrodes:

The system base has two stainless-steel rectangular foot-pad electrodes fastened to a metal platform set on force transducers for weight measurement; each of the extremity hand-grip electrodes has an anterior and posterior portion. The eight electrodes are connected to a digital circuit board that electronically switches the electrical circuit under study. The electrode configuration protocol allows the body mass readings for the right arm, left arm, trunk, right leg and left leg. This device provided further data on weight, segmental and total fat percentage (FP), fat mass (FM), fat-free mass (FFM), BMI, total body water and basal metabolic rate. The use of this system in the determination of body composition has been validated against DXA method in healthy adults [20].

Precision study

All clinical measurements were performed by trained personnel. The in vivo coefficient of variation (CV %) for the 30 duplicated measurements from the BIA measurements was calculated. The CV was 0.91%, and the inter-observer CV was 1.02%.

Statistics

Data were analyzed using the SPSS software (version 19; SPSS Institute, Chicago, USA). All body composition parameters and descriptive statistics in this study were normally distributed. The relationship between the measured variables (when appropriate) was tested using a regression analysis (the curve estimation procedure); the BMI was set as the dependent variable. A stepwise regression analysis was used to determine the best predictors of BMI. Statistical significance was set at p ≤ 0.05. To provide estimates for the whole population, totals were weighted according to the age distribution of the women over 18 in the Extremadura region [21].

Results

Prevalence of overweight and obesity

According to the BMI criteria, the prevalence of overweight in Extremadura’s Spanish women was 38.4% (95% CI 36.9-39.9%; n=1,539), and the prevalence of obesity was of 24.6% (95% CI 23.3-26.0%; n=989). A total of 1,485 women (36.9%; 95% CI 35.4-38.4%) were considered normal according to the BMI criteria. Weighted prevalence for overweight and obesity according to the age distribution of the women over 18 in the Extremadura region were 34.44% (95% CI 34.23-34.67%) and 21.66% (95% CI 21.48-21.86%) respectively. The overweight and obese women were stratified by age groups as shown in Table 1. Normal weight women in the population were more likely to be between 15 and 44 years of age (more than the 50% of the sample). In the 45-54 year age group, overweight women were more common than normal weight women (42.1% n=613 vs. 39.6% n=577). Obese women were significantly (p<0.05) more common in the 55-64 year age group (36.1% 95% CI 33-38%; n=387).

Body composition analysis

A total of 2,224 women were included in the body composition analysis. The BIA cohort was aged 54.45±10.20 years (18-85 years) and the non-BIA cohort 53.42±11.99 years (18-87 years) (p=0.0003). No statistically significant differences, were found in the BMI between the two groups (p>0.05). The results for body composition in the total sample based on the anatomical segment are shown in Table 2. In the BIA cohort, bivariate correlation between BMI and FP was lower (r=0.83, p<0.0001) than the one between BMI and weight (r=0.89, p<0.0001) but higher than the correlation between weight and FP (r=0.80, p<0.0001).

When participants in the study were categorized by BMI group (normal, overweight or obesity) some changes were observed in the aforementioned correlations. Correlation between BMI and FP were lower in all the groups r=0.67, r=0.44 and r=0.62 respectively for normal, overweight and obesity (p<0.0001 in all cases). Contrary to the observed in the whole BIA cohort, BMI correlation with weight was lower in all the groups, r=0.60, r=0.50 and r=0.77 for normal, overweight and obesity groups (p<0.0001 in all cases). Correlation between weight and FM were lower than in the whole cohort (r=0.60 in the normal group, r=0.43 in the overweight group and r=0.62 in the obesity group, p<0.0001 in all cases).

Age-specific percentile values

Age-specific percentile values for the 10th, 25th, 50th, 75th and 90th percentiles are presented in Tables 3, 4, 5,6, 7, and 8 (as supplemental files) and in Figure 1 and 2. The FM gradually increased between 18 and 74 years of age and decreased thereafter. The FM in the right leg also was increased between 65 and 74 years of age (5.69 ± 1.4 Kg) and decreased thereafter. A similar result was found in the left leg, where the FM increased between 65 and 74 years of age (5.66 ± 1.4 kg) and decreased to 5.44 ± 1.32 thereafter. In the upper extremities, 90% of the studied samples had a FP ranging from 22.43 to 49.8%. The FM in the right arm increased until 44 years of age (1.22 ± 2.6 Kg), decreased to 1.19 ± 0.5 Kg between 45 and 54 years and increased again to 1.54 ± 0.63 from 65 to 74 years of age. The FM in the left arm increased constantly until it reached a peak of 1.63 ± 0.6 Kg between 65 and 74 years of age; it decreased thereafter. We observed that the FM increased in the trunk throughout life (peaks at 13.27±3.8 Kg) until the subject reached between 65 and 74 years of age. FFM increased until the 44 years (42.42 ± 4.17 Kg), and decreased thereafter. The lower value of FFM was found in the group of women over 74 years with 39.78 ± 4.62 Kg. In this group, a 10% of the women presented a FFM below 17.3 Kg (Figure 2).

Discussion

According to the internationally accepted thresholds for BMI, approximately 63% of the population exceeded the recommended threshold for healthy body habitus. Our estimated prevalence for obesity (24.6% 95% CI 23.3-26%) exceeded the prevalence estimated by the nationwide DORICA study (17.5% for obesity and 32.9% for overweight among women between 25 and 60 years of age) [22] and exceeded the estimated obesity prevalence calculated by the Spanish Health Ministry in their last nationwide health survey (15.3%) [23]. A recent local study in our area, based in the direct measurement of BMI in a sample of 1,515 women aged 51.2±14.2 years have reported higher figures for overweight (33.2%) and obesity (32.6%) [24]. The prevalence of obesity and overweight observed in the region of Extremadura is also higher than the reported in a large cohort of 1,726 adult women (aged 39.6±11.3 years) in the Balearic Islands of 22.8% overweight and doubles the prevalence of obesity (10.9%) [25].

The analysis of the available data [22–39] indicates that the prevalence of obesity in Spain is dependent on geography. . Our results are similar to those obtained in previous studies and add to the current knowledge about the prevalence of obese and overweight women from the south of Spain, , which most likely reflects the common habits of our population and the similar environment.

Determining the BMI does not require sophisticated equipment and is easy to calculate. As previously indicated, the Spanish Health Ministry proposed an estimated prevalence of obesity of 15.3% in the women of our area. This was the result of a telephone-based survey that was developed in 2009. It has been widely described that a self-reported BMI might be biased because of the tendency to overestimate height and underestimate weight, which underrates BMI by up to 6.7% [40]. Our study supports the hypothesis that the prevalence of obesity in our area is one of the highest in Spain. Nevertheless, comparisons need to be performed with studies that have effectively addressed the BMI and not with self-reported data.

Lopez et al, [25] reported recently similar correlation coefficients than the observed in our study. Between BMI and FP a r=0.80 (p<0.001) was observed. Slight higher correlation was reported for BMI and weight (r=0.89, p<0.001). However, lower correlation was observed for weight and FP (r=0.72, p<0.001). With BIA Lopez et al reported a FP of 31.6±7.1% also lower than the observed in our sample (36.32±6.35%). This could be explained by the fact that Lopez et al cohort was younger than ours, and an age-dependent increase in adiposity has been described in women [25]. An older cohort was studied by Montero-Lopez et al in the closer regions of Toledo and Madrid, in central Spain. Based on a cohort of 287 women aged 54-75 y and after BIA analysis, FP was established in a 38.20±7.14% more similar to the observed in our cohort [41].

Body composition in a cohort of 716 Spanish women, measured by DXA, has been also reported. Correlation coefficient reported by this technique between BMI and FP was of r=0.77 (p<0.001) lower than the described in our study. Overall in the age ranges of 21-80 years Aguado-Henche et al, [42] reported higher FP than the observed in our cohort. We think that these confirm the shown in validation studies: BIA may underestimate FP systematically compared with both the underwater weighing method (the gold standard for body composition) [43] and DXA [19]. However, we have observed an age based FP distribution with BIA similar than the observed in the study of Aguado-Henche. In their study, FP was increased from 35 to 55 y stabilizing thereafter until de age of 70. A similar result was observed in our cohort with a continuous increase from 32.27±6.24% at 35-44 years to 38.16±6.01% at 55-64 years and then 39.69±5.14% for 65-74 years and 39.26±4.88% for the >74 years women.

Previously published BIA studies in Spain [25,41,44] are generally in line with the findings reported here, although these analyses did not match on age and sample size with our study. Our estimates, with a large population-based sample and BIA, confirm previous results with smaller samples and provide reliable quantitative estimates of the FP, FM and FFM in Spanish women from our area. On the other hand, it is difficult to compare the numbers of DXA in the study of Aguado-Henche et al, with those of the BIA that we present because 2 different methods were used to measure.

By using a representative sample of Spanish adult women, we described the segmental distributions of FP, FM and FFM and provided their quartile distribution according to age. The estimated percentiles cutoffs of FP, FM and FFM at commonly used percentile by age in women may provide an effective reference in epidemiological settings and public health services. The data provided in this article could be used to estimate a woman’s absolute FP, FM and FMM according to her age. The dilemma with percentiles is that they are population specific.

We acknowledge that our study has several limitations. First, we used BIA as the reference methods for body composition analysis. It has been reported that the use of BIA may underestimate the FP as much as 12% in the population at the lower ranges of FP, and it may overestimate FP even with 8% in the upper range of FM% [19]. Second, the provided percentiles are population specific, and generalization to the Spanish female population may not be appropriate. Further studies are necessary in other Spanish regions for comparative purposes. In contrast, the strengths of our analysis include the use of the largest sample to date, and the use of an entirely harmless and cheap technique.

This is the first study that has addressed the body composition parameters in a large cohort of Extremadura’s Spanish women. We have provided data for the elaboration of reference curves that might be of help from an epidemiological point of view. BIA is actually recognized as a low-cost, safe, accurate and reproducible method for body composition analysis [45]. Overall, our results indicate that the use of BIA provides an opportunity to evaluate the utility of body-composition estimates as predictors of the risk of chronic diseases. These results are acceptable for epidemiological studies, but some inevitable inaccuracy in the data must be expected