Relationships between body composition and pulmonary function in a community-dwelling population in Japan

Pulmonary diseases, including chronic obstructive pulmonary disease (COPD), are major chronic diseases that result in decreased pulmonary function. Relationships between body composition and pulmonary function have been reported. However, few epidemiological studies have used the visceral fat area (VFA) to measure body composition. This study aimed to examine the relationship between body composition and pulmonary function. A cross-sectional study was conducted between 2015 and 2016, using data obtained from 1,287 residents aged between 19 and 91 years living in the Iwaki area of Hirosaki City, a rural region in Aomori Prefecture, Japan. Pulmonary function was evaluated using the forced vital capacity (FVC) as a percentage of the predicted value (predicted FVC%) and the ratio of forced expiratory volume in one second (FEV1) to FVC. The measurements for evaluating body composition included the body fat percentage (BFP) of the whole body and trunk, skeletal muscle index (SMI), body mass index (BMI), VFA, waist circumference (WC) at the navel level, and waist-to-hip ratio (WHR). To adjust for potential confounders, Spearman’s partial correlation analysis was used to examine the relationship between the measurements of body composition and pulmonary function. There were significant correlations between the predicted FVC% and the following parameters: BFP (whole body and trunk) in younger males; SMI in older males; WC, VFA, BMI, and SMI in younger females; and BFP (whole body and trunk) and VFA in older females. Contrastingly, WC and VFA in younger males and WC in younger females were correlated with the FEV1/FVC ratio. VFA was correlated with the FEV1/FVC ratio in younger males and predicted FVC% in older females. These findings suggest that visceral fat accumulation may increase the development of obstructive pulmonary disease in young males and accelerate the decline of pulmonary function (predicted FVC%) in older females.


Introduction
Smoking status [1], aging [2], decreased muscle mass [3], and obesity [4] impair pulmonary function to varying degrees and may induce severe respiratory diseases, including chronic obstructive pulmonary disease (COPD) [5]. Intra-abdominal fat accumulation may impede diaphragm descent during inspiration, affecting several spirometry variables [6]. Moreover, fat deposition on the chest wall may impede rib cage expansion and excursion [7]. In addition, metabolic syndrome contributes to obstructive pulmonary disease and is influenced by glucose tolerance and inflammatory cytokines [8]. Therefore, the visceral fat area (VFA) is likely related to pulmonary function. Recently, computed tomography (CT) [9], magnetic resonance imaging [10], and dual-energy X-ray absorptiometry [11] have been demonstrated as effective modalities for the relatively accurate measurement of total body fat and visceral fat. However, these modalities have limited applications in large-scale and general population surveys because of radiation exposure-related problems, technical complexity, and high cost.
Increasing attention has been given to impedance methods that are unaffected by the aforementioned problems. In particular, the eight-polar bioelectrical impedance analysis [12] allows accurate measurement of each site. However, considering the increasingly important role of visceral fat in metabolic syndrome, specialized techniques for measuring intra-abdominal fat are required. Consequently, a method for measuring visceral fat was developed, which used both impedance methods in the anterior and posterior axial directions of the trunk [13] and dual bioelectrical impedance analysis [14]. The results from both methods were highly correlated with CT measurements. However, there is a need for further study in the healthy general population and young people to identify the relationship between body fat (including intraabdominal fat) and pulmonary function, with an emphasis on prophylaxis.
The improved measurement accuracy of the impedance method, which measures visceral and subcutaneous fat separately, has allowed its use in body fat measurements for epidemiological studies.
The features of this study are as follows: 1. A relatively large, healthy population was studied.
2. Visceral fat was measured using the impedance method.
3. Various confounding factors that affect pulmonary function and body composition (including body fat) were measured. 4. Homeostasis model assessment of insulin resistance (HOMA-IR) and interleukin-6 (IL-6) levels were determined to elucidate the mechanism underlying the prevention and improvement of obstructive pulmonary diseases.
To our knowledge, there has been no report on the association between pulmonary function and body composition (including VFA) in the healthy general population. The mechanisms underlying prevention and improvement of pulmonary function and body composition remain unclear. Elucidating the association of body composition measurements, specifically the VFA and obesity criteria, with pulmonary function and determining the underlying mechanism may provide basic data for the prevention and improvement of obstructive pulmonary diseases. Therefore, we aimed to examine the relationship between body composition and pulmonary function. employed by commercial company: Health Care Food Research Laboratories, Kao Corporation, Tokyo, Japan. However, the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The funding organization of YK only provided financial support in the form of YK's salaries and research materials.
Competing interests: The author Y.K. received funding from Health Care Food Research Laboratories, Kao Corporation, Tokyo, Japan, a commercial company, in the form of salary and research materials. There are no patents, products in development or marketed products to declare. This does not alter our adherence to all the PLOS ONE policies on sharing data and materials. 2015 and 2016. When participants were included both in 2015 and 2016, their 2015 data was adopted. We excluded participants who reported a previous diagnosis of heart disease, stroke, or cancer (n = 153) and those who did not fully complete the pulmonary function/body composition measurements or the questionnaire, which included medical history and lifestyle behavioral questions (n = 10). Finally, 1,287 participants (498 males and 789 females) were included for analysis. For further analysis of the association between VFA and the ratio of forced expiratory volume in one second (FEV 1 ) to forced vital capacity (FVC) in the younger male group, 383 participants in the younger male group were selected from the 1287 participants. Among these 383 participants (aged 19-64 years), 9 with diabetes (fasting blood glucose � 126 mg/dL [15]) and 3 with missing data were excluded. Consequently, 371 people were included in the analysis. All participants provided informed written consent; further, this study was conducted in accordance with the Declaration of Helsinki. Ethical approval was provided by the Ethics Committee of the Hirosaki University Graduate School of Medicine (approval numbers: 2014-377, 2016-028).

Survey items
Questionnaire investigation. Each participant completed a self-administered questionnaire regarding sex, age, and smoking status (current smoker, ex-smoker, non-smoker).
Body weight, body fat percentage, and skeletal muscle mass. A body composition analyzer (MC-190; Tanita Corp., Tokyo, Japan) [12], using the multi-frequency bioelectrical impedance method, was used to measure body mass (kg), body fat percentage (%), and skeletal muscle mass (kg). The body mass index (BMI) was calculated by inputting the measured height to this body composition analyzer. We used BMI � 25 kg/m 2 as the obesity criterion [16]. The skeletal muscle mass was evaluated using the skeletal muscle index (SMI), calculated as the appendicular muscle mass divided by the height squared [17].
VFA. The visceral fat-measuring apparatus (EW-FA90; Panasonic, Osaka, Japan), which used the bioelectrical impedance analysis method [13], was used for waist and VFA measurements. We used a VFA � 100 cm 2 as the visceral fat obesity criterion [16]. The waist circumference (WC) was simultaneously measured at the navel level using this device. We used � 85 cm in males and � 90 cm in females as the obesity criteria [16].
Waist-to-hip ratio. The waist-to-hip ratio (WHR) was calculated as WC divided by hip circumference. We used � 0.90 in males and � 0.85 in females as the obesity criteria of WHR [18].
Pulmonary function. FVC and the FEV 1 /FVC ratio measurements were obtained using a multi-functional spirometer (HI-801; CHEST MI, Inc., Tokyo, Japan) [19]. Pulmonary function was measured following the guidelines of the Japanese Respiratory Society, which are in accordance with those of the American Thoracic Society/European Respiraroty Society [20]. Predicted FVC% was calculated as FVC divided by the predicted FVC estimated using the LMS method [21] as follows: predicted FVC ¼ expðÀ 8:8877 þ 2:1494 � lnðhÞ À 0:1891 lnðaÞ þ mÀ sÞ where a = age in years; h = height in cm; ln() = natural log transformation; m-s = M-spline by age.
Blood analysis data. a) HOMA-IR. HOMA-IR was performed by homeostasis model assessment based on fasting blood glucose and insulin levels, with levels � 2.5 being considered abnormal [22]. b) IL-6. Serum IL-6 concentration was measured by chemiluminescent enzyme immunoassay using a commercially available kit (LSI Medience, Tokyo, Japan); values > 4 pg/ml were considered abnormal.

Statistical analysis
All participants were categorized according to sex and age (19-64 years [the younger group] and � 65 years [the older group]). Descriptive statistics are presented as median, 25 th and 75 th percentile, or percentage. The Mann-Whitney U-test was applied to compare measurement variables between the obesity and age groups. The Kruskal-Wallis test was applied to compare measurement variables between the sum of obesity criteria (0, 1, and 2-4) and pulmonary function. Within-group differences were evaluated with post hoc Bonferroni adjustment. Spearman's partial correlation analysis was used to examine the relationship between body composition measurements and pulmonary function. This analysis was performed with adjustment for smoking status, age (only for the FEV 1 /FVC ratio), and SMI (except when analyzing the SMI). Because the predicted FVC calculated from the LMS method uses age as a predictor, age was not adjusted for in the analysis of the predicted FVC% value. Smoking status was categorized as current-smoker, ex-smoker, and non-smoker, and was included as a dummy variable. Between-group differences in the smoking status were examined using the chi-square test.
Additionally, Spearman's partial correlation analysis was used to examine the relationships among the measurements of HOMA-IR, IL-6, and the FEV 1 /FVC ratio with adjustment for smoking status, age, and SMI. These variables were converted into dummy variables and entered in the analysis as adjustment variables. The Mann-Whitney U test was applied to compare the FEV 1 /FVC ratio between the normal and abnormal values of HOMA-IR and IL-6 and to compare HOMA-IR and IL-6 between the normal and abnormal values of VFA. In addition, Spearman's rank correlation analysis was used to examine the relationship between VFA and IL-6. Statistical significance was set at p < 0.05. IBM SPSS Statistics for Windows, version 25.0 (IBM Corp., Armonk, N.Y., USA) was used for analysis. Table 1 shows the characteristics of the male participants. Age (p < 0.001), WHR (p < 0.001), body fat percent (p = 0.022), and body fat percent on the trunk (p = 0.022) were significantly higher in the older group than in the younger group; the opposite was the case with their height (p < 0.001), weight (p < 0.001), predicted FVC% values (p = 0.001), FEV 1 /FVC ratio (p < 0.001), and SMI (p < 0.001). Furthermore, there was a significantly higher frequency of non-smokers in the older group. Table 2 shows the characteristics of the female participants. Age (p < 0.001), WC (p < 0.001), VFA (p < 0.001), WHR (p < 0.001), body fat percent (p < 0.001), and body fat percent on the trunk (p < 0.001) were significantly higher in the older group than in the younger group; the opposite was the case with their height (p < 0.001), weight (p = 0.003), FEV 1 / FVC ratio (p < 0.001), and SMI (p < 0.001). Similar to the male participants, there was a significantly higher frequency of non-smokers in the older group. Tables 3-6 show the relationship between the presence of obesity criteria outlier and pulmonary function. In the younger groups of both sexes, there was a significantly stronger relationship between the FEV 1 /FVC ratio and VFA in the non-obese groups than in the obesity groups (males, p = 0.002; females, p = 0.044). Further, more younger males in the non-obese group met the 0 obesity criterion than the 2 obesity criterion (p < 0.001). There was no relationship between body composition and pulmonary function in the older male groups. In the older female groups, the relationships of the predicted FVC% with WC (p = 0.003), VFA (p = 0.002), and BMI (p = 0.023) were significantly higher in the non-obese group than in the obesity group. Table 7 shows the relationship between the body composition index and pulmonary function.

Relationship between the body composition index and pulmonary function
In the younger male group, there was a significantly negative correlation between the predicted FVC% and body composition index (r = −0.102, p = 0.047 for body fat percentage; r = −0.108, p = 0.035 for body fat percentage of the trunk), as well as a significantly negative correlation    Relationships between HOMA-IR, IL-6, and the FEV 1 /FVC ratio in the younger male group Tables 8 and 9 show the relationships between HOMA-IR, IL-6, and the FEV 1 /FVC ratio in the younger male group. There was no significant correlation between HOMA-IR, IL-6, and the FEV 1 /FVC ratio. Individuals with abnormal IL-6 levels had a significantly lower FEV 1 / FVC ratio than those with normal IL-6 levels (p = 0.045). However, there was no difference in the FEV 1 /FVC ratio between those with and without abnormal glucose tolerance. Spearman's partial correlation analysis between body composition and pulmonary function was used.
This analysis was adjusted for smoking status, age (only for the FEV 1 /FVC ratio), and SMI (except when analyzing the SMI). Smoking status was categorized as currentsmoker, ex-smoker, and non-smoker; these were included as dummy variables.
Abbreviations: Predicted FVC%, the FVC as a percentage of the predicted value.
https://doi.org/10.1371/journal.pone.0242308.t007 Spearman's partial correlation analysis was used to determine the relationship between body composition and pulmonary function.
This analysis was adjusted for confounders, including smoking status and age. Smoking status was categorized as current-smoker, ex-smoker, and non-smoker, which were included as dummy variables. HOMA-IR, homeostasis model assessment of insulin resistance; IL-6, interleukin-6.
https://doi.org/10.1371/journal.pone.0242308.t008 Tables 10 and 11 show the relationship of VFA with HOMA-IR and IL-6 in the younger male group. VFA showed a significant positive correlation with HOMA-IR (r = 0.570, p< 0.001) and IL-6 (r = 0.254, p< 0.001). In addition, there were significantly more individuals with abnormal values for VFA than those without (p< 0.001 for both HOMA-IR and IL-6).

Discussion
This study examined the association between body composition and pulmonary function. In the younger age groups of both sexes, the FEV 1 /FVC ratio was significantly higher in the normal VFA group than in the obesity group (Table 3). However, after adjusting for confounding factors, the younger male group showed a significant negative correlation between the FEV 1 / FVC ratio and VFA only (Table 7). Because the FEV 1 /FVC ratio is used as a diagnostic criterion for obstructive pulmonary disease [23], VFA accumulation may increase the risk of obstructive pulmonary disease in young males. The relationship between the FEV 1 /FVC ratio and VFA has been reported in studies conducted in Korea [24] and the Netherlands [25], with none of the studies reporting a significant relationship. Further studies are needed to confirm these findings. On the contrary, in the older female group, the predicted FVC% was significantly higher in the normal WC, VFA, and BMI groups than in the obesity group (Table 6). However, the older female group showed a significant negative correlation of the predicted FVC% with the body fat percentage, body fat percentage of the trunk, and VFA, after adjusting for confounding factors. The study results suggest that the increase of VFA may accelerate the decline of pulmonary function (predicted FVC%) in older females.
Metabolic syndrome affects the FEV 1 /FVC ratio through various factors such as increased glucose tolerance and inflammatory cytokines [8]. The present study showed no correlation among pulmonary function, glucose tolerance, and inflammatory cytokines. However, the FEV 1 /FVC ratio was significantly lower in young men with abnormal IL-6 levels than in those with normal IL-6 levels. In addition, there was a positive correlation between VFA and IL-6 concentration in the same group. Further, IL-6 concentration was significantly higher in individuals with abnormal VFA than in those without. Therefore, it is likely that IL-6 was positively correlated with VFA in younger men, which may have resulted in the decrease in the FEV 1 /FVC ratio. Previous studies compared visceral fat and IL-6 in the obstructive and nonobstructive pulmonary disease groups and showed that the former had significantly greater visceral fat and IL-6 levels than the latter [26,27]. In addition, these studies compared VFA and IL-6 and analyzed the independent associations of obstructive pulmonary disease status and VFA on iL-6 plasma concentrations. The present findings are consistent with previous findings and suggest that excessive abdominal visceral fat contributes to increased IL-6 levels, which causes obstructive pulmonary disease in older adults [27]. In women, decreased estrogen and progesterone levels with menopause lead to increased visceral fat [28]. Furthermore, VFA and VFA/subcutaneous fat area after menopause are significantly higher than those before menopause [29]. In this study, only elderly women showed an association between VFA and predicted FVC%. The increase in visceral fat in older women in the menopausal group may have contributed to decreased respiratory compliance and consequently affected the association between VFA and predicted FVC%.
A study on Western participants reported negative correlations between WHR, BMI, and pulmonary function [30]. In this study, there was no relationship between WHR and pulmonary function; moreover, there was a relationship between BMI and pulmonary function in females only. This result was influenced by the physical characteristics of the Japanese population. The percentage of male and female Americans with BMI � 30 kg/m 2 is 35.0% and 40.4%, respectively [31], while the corresponding values for Japanese are 4.5% and 3.9% [32]. The corresponding values in the present study were 3.8% and 3.0%. It has been shown that FVC and FEV 1 significantly decrease when the BMI and WHR are � 30 and � 1.01, respectively [33]. Further, it has been suggested that the effect on pulmonary function remains small except in cases with extreme obesity [34]. Since the participants in this study had less extreme obesity, it could have attributed to the findings of no relationships of BMI and WHR with pulmonary function.
This study showed a strong association between WC, VFA, and pulmonary function. It has been reported that the VFA to subcutaneous fat ratio is higher in the Japanese than in other races [35,36]. Therefore, the Japanese may be more affected by diseases related to visceral fat than other races. It has been suggested that VFA measurement is important as a disease marker for pulmonary disease in the Japanese. The effects of obesity on pulmonary function are more associated with FVC than with FEV 1 [37]. Furthermore, VFA is positively correlated with intra-abdominal visceral fat volume [38], with numerous studies examining the relationship between FVC and FEV 1 (VFA, FVC, and FEV 1 showed negative correlations) [25,[39][40][41]. Values are the median and the 25 th and 75 th percentiles. The p-value was obtained using the Mann-Whitney U-test between the applicable and not applicable groups, in terms of abnormal visceral fat area.
HOMA-IR, homeostasis model assessment of insulin resistance; IL-6, interleukin-6. https://doi.org/10.1371/journal.pone.0242308.t011 However, there have been few reports on the relationship between VFA and the FEV 1 /FVC ratio. In this study, VFA in young males was related to the FEV 1 /FVC ratio but did not predict FVC%. The present study examined the relationship between body composition and pulmonary function in the general population of the Aomori Prefecture lacking pulmonary diseases. There was a weak but significant association between VFA and the FEV 1 /FVC ratio in the younger male group. The possibility that obesity prevention may improve pulmonary function from such a period of no apparent disease is immensely important from an epidemiological perspective.
There were some limitations to the present study. First, because this was a cross-sectional study, we could not prove causality. Longitudinal studies are needed to reveal the causal relationship between body composition and pulmonary function. Second, all the participants were volunteers; therefore, they might have been more interested in their health status and could have been healthier than the general population. Nevertheless, even with these limitations, the present study measured the VFA in a sample size relatively larger than that of previous studies. Our study provided basic evidence regarding the relationships between body composition and pulmonary function.

Conclusions
This study, which examined the association between body composition and pulmonary function, provides basic data for the prevention and improvement of obstructive pulmonary diseases. The FEV 1 /FVC ratio was significantly higher in the younger male group with normal VFA than in those with obesity. Moreover, there was a significantly negative correlation of the FEV 1 /FVC ratio with WC and VFA after adjusting for confounding factors in the same group. These results suggest that increased VFA could promote the development of obstructive pulmonary disease. By contrast, in the older female group, the predicted FVC% was significantly higher in the normal WC, VFA, and BMI groups than in the obesity group. In addition, there was a significantly negative correlation of the predicted FVC% with body fat percentage, body fat percentage of the trunk, and VFA. The study results suggest that increased VFA may accelerate the decline of pulmonary function (predicted FVC%) in the older female group. In addition, IL-6 may be related to a decrease in the FEV 1 /FVC ratio associated with visceral fat accumulation. Furthermore, VFA measurement may be an important disease marker for pulmonary disease among the Japanese population.