Juntendo Medical Journal

Objective: This study examined the prevalence of and factors affecting normal weight obesity (NWO) among women aged under 40 years old. NWO was having a normal BMI but a high body-fat percentage (BFP). Materials and Methods: We recruited 399 participants aged 18­39 during a health checkup and surveyed them about lifestyle patterns, height, weight, body composition, and girth and visceral fat area. Participants were divided into three. Those with a BMI of <25 and a BFP of <30 were classified as non-obese, a BMI of <25 and a BFP of ≥30 as NWO, a BMI of ≥25 and a BFP of ≥30 as obese. The analysis was conducted on NWO (97) and non-obese (262) groups. Results: NWO prevalence ranged from 20% to 30% across the three age groups. NWO participants had lower values for segmental lean body mass, particularly in lower limb, and higher values for abdominal girth and visceral fat area. Among participants aged 30­39, those with NWO had higher values for triglycerides and LDL cholesterol and lower values for HDL cholesterol. Regarding lifestyle factors, among those aged 30­39, NWO was associated with past weight fluctuation, respectively. In the latter age group, a gain of more than 10 kg since the age of 20 was associated with a 13-fold increase in the odds for NWO. Conclusions: Results suggest that for women under 40 years, detecting NWO early and guidance on effective lifestyle and exercise are necessary.


Introduction
According to World Health Organization (WHO), "Worldwide obesity has tripled since 1975, and by 2016, over 1.9 billion adults were overweight, of which 650 million are obese" 1) . Obesity is an excessive accumulation of body fats and is usually determined by a personʼs body mass index (BMI), defined as"a simple index of weightfor-height commonly used to classify overweight and obesity in adults" 1) . Moreover, "for adults, WHO defines overweight and obesity as follows: over-weight is a BMI greater than or equal to 25; and obesity is a BMI greater than or equal to 30" 1) .
The Organization for Economic Co-operation and Development (OECD) reported obesity rates of different countries, highlighting obesity as prevalent in these countries 2) . In the U.S., for example, over 30% of adult men and women (aged 15 and over) had a BMI greater than or equal to 30 2) . Japan Society for the Study of Obesity (JASSO) defined obesity as BMI greater than or equal to 25, a threshold lower than WHOʼs 30. Japanʼs National Health and Nutrition Survey found that almost 30% 1 of men and 20% of women in Japan had a BMI of 25 or greater, whereas only 3% had 30 or greater. Despite Japanʼs relatively low obesity rates, it is at par with other OECD countries in terms of prevalence of obesity-related disorders, such as type 2 diabetes. JASSOʼs low threshold for obesity indicates that prevalence and onset rates for obesity-related disorders are high, despite obesity rates being low 3) .
Although BMI is the prevailing measure of fat levels, it has a limitation. As it is a weight-forheight index, it is affected by muscle mass and fluid retention. Therefore, BMI does not fully reflect body fat percentage (BFP). To accurately gauge BFP, it is necessary to ascertain fat accumulation using measures other than weight and height.
Studies in the 1990s noted cases where a person has high BFP despite normal BMI ─ termed"metabolically obese normal weight"or"normal weight obesity"(NWO) 4) 5) . Takeda et al. (2017) reported that NWO was present in 4%40% of female adolescents in Japan 6) . Franco et al. (2016) reported that almost 30 million Americans were estimated to have NWO 7) . In 2007, a Japanese study reported that NWO is associated with lifestyle diseases similar to regular ( "overweight" ) obesity, such as arteriosclerosis and diabetes 8) . Other Japanese studies have noted that people with NWO do not exercise enough or dislike physical activity and exhibit unhealthy dietary habits 9) 10) .
Literature has underscored the need to identify NWO. However, obesity research has largely ignored individuals with NWO, as they fall within the normal BMI range. Consequently, there is a paucity of detailed findings on NWO 9) 10) . The few studies examining NWO were conducted on a small and age-limited sample (female college students) 5) 6) 8) 10)-13) . Moreover, unlike regular obesity, there is no internationally accepted definition of NWO; meaning, the diagnostic criteria vary among different studies 7) 14) 15) .
In 2008, the Japanese Government launched a program on health checkups that focus on visceral fat and BMI. This program makes it compulsory to monitor visceral fat in adults over the age of 40 by measuring abdominal girth. However, girth measurements are not compulsory for those below 40; for these individuals, obesity is still based on only BMI. Accordingly, the present study was con-ducted on women under the age of 40 and thus were not yet subject to health checkups. Findings on NWO are expected to enable the identification of factors affecting susceptibility to the onset of obesity. Understanding these factors will guide efforts towards preventative measures.

Materials and Methods
The study was conducted between July 2016 and February 2017 on 399 women aged 1839 from areas across Japanʼs Mie Prefecture, which has a mild climate and is located almost in the center of Japan. The sites included urban cities and local cities. The participants were community residents that got the health checkups, which are legally required, performed regularly, and include physical measurement, medical examination, and a laboratory test.
A questionnaire survey about lifestyle patterns and the physical data such as height, weight, body composition, abdominal girth, and visceral fat area (VFA) was administered.
Body composition was measured in several ways ─ hydrostatic underwater weighing, air displacement plethysmography (ADP), dual-energy X-ray absorptiometry (DXA), and bioelectrical impedance analysis (BIA) 16) . In this study, we used BIA on the basis that it would minimally encumber participants, involve minimal exposure, and is a straightforward and inexpensive method. Thus, eight electrodes were fitted to hands and feet of participants who were standing. A weak alternating current was passed through the body at multiple frequencies. The investigator calculated impedance (electrical resistance) in the limbs and trunk, which reflected the concentration of electrolytes in these body segments. The method was premised on the fact that the bodyʼs water, which contains electrolytes, is mostly stored in muscles and other non-fatty (lean) areas. Fatty areas, which contain relatively low levels of water, are less conductive than lean areas are, and therefore cause greater impedance 16) . We used three eight-polar BIA devices: InBody 430, InBody 470, and InBody 770 (InBody Japan, Tokyo). These devices were used at the four sites. InBody 430 was used at two sites; InBody 470 was use at one site; and InBody 770 at another site. However, we did not break down results by device because the devices were Nakanishi Y, et al: Status of normal weight obesity among Japanese women under 40 years old largely identical in terms of measuring method and items measured 17) . Typically, abdominal girth is determined using a measuring tape during a health checkup, whereas VFA is measured using a CT scan. In this study, a CT scan was not feasible and may have alarmed participants worried about x-ray exposure. Visceral fat and abdominal girth were measured using a medically approved visceral fat meter, EW-FA 90 (Panasonic, Osaka), which, measures abdominal BIA and is portable and easy to use; its measurements correlate with a CT scan 18) 19) .
Participantsʼ lifestyle patterns were determined based on responses to 12 items likely to be related to obesity or a lifestyle disease, extracted from a 2016 20) .
This study involved laboratory data comparisons; any data collected in health checkups was used with consent from participants.
We explained the purpose, method, and Participants were briefed about the studyʼs purpose and method, and made aware that participation was voluntary. Written informed consent was obtained, including consent to participate and publish findings. The research protocol was reviewed and approved by the research ethics committee of the Graduate School of Health and Sports Science, Juntendo University (#27-54). For participants aged 18 or 19 years, parental/guardian consent was obtained from in writing.

Analysis
Of 399 participants, 397 were included in the analysis, after excluding those with missing data. Of the 397 included in the analysis, 48 were aged 1819 years, 99 were aged 2029 years, and 250 were aged 3039 years.
Regarding NWO prevalence in the sample, in view of the literature 6) 8) 10) 12) 13) , NWO was defined in this study as a BMI of < 25 and a body-fat percentage (BFP) of ≥30%. Participants were grouped into three categories based on their BMI and BFP values (as measured by the InBody devices). Those with a BMI of < 25 and a BFP of < 30% were classified as non-obese; a BMI of < 25 and a BFP of ≥30%, as NWO; a BMI of ≥25 and a BFP of ≥30%, as obese. The obese group covered 38 participants who were excluded from the analysis due to small group size would have reduced the statistical power. The analysis was then conducted on the two remaining groups: 97 NWO and 262 non-obese participants. Distributions of the measured items and health record items were subjected to a Kolmogorov-Smirnov test, revealing certain items that lacked a Gaussian distribution. Accordingly, a Mann-Whitney U test was performed.
To determine factors related to NWO, a chisquared test was performed, comparing NWO to the 12 survey items on lifestyle. Next, a binary logistic regression was performed; dependent variables were NWO and non-obesity, whereas independent variables were survey items that the chi-squared test revealed to be related to NWO.
Each analysis was performed separately on three age groups (1819, 2029, 3039) to determine age-based differences. Statistics were computed using SPSS (IBM Japan, Tokyo). Statistical significance was determined with a threshold of p < 0.05.

Factors related to NWO 1) Measurements
The U test revealed that NWO participants, regardless of age, had significantly greater abdominal girth and VFA compared to their non-obese counterparts (p < 0.001).
Given that muscle mass varies widely by constitution (height, weight, BMI) 21) , we compared each participantʼs segmental (upper and lower limbs, trunk) lean body mass (LBM) against the optimal segmental LBM for that personʼs current body weight, scaling the latter at 100%. Among participants aged 2029 or 3039, NWO participants fell significantly lower of their LBM benchmark than their non-obese counterparts with respect to all body segments. However, in the 2029 age range, intergroup difference for the upper limbs was no lower than 5%. Among participants aged 1819, intergroup difference was significant for the trunk (p < 0.01) and lower limbs (p < 0.001) but not for the upper limbs (see Table-2, sections 1, 2, 3).

2) Laboratory data
The medical check records contained laboratory data on 329 of 359 participants. Such laboratory data were analyzed. There was no significant intergroup difference in laboratory data among participants aged 1819 or 2029. Among those aged 3039, however, the two groups differed significantly with respect to three laboratory test items; compared with their counterparts in the same age range, NWO participants had significantly higher results for triglycerides (p < 0.001) and LDL cholesterol (p < 0.05), and significantly lower results for HDL cholesterol (p < 0.05). There were no significant intergroup differences in HbA1c or adiponectin. (see Table-2, sections 1, 2, and 3).

3) Relationship to obesity
Despite the small size of the obesity group, to confirm the difference between NWO and obesity, we also tried a U test between both groups. The U test revealed that obese participants, regardless of age, had greater abdominal girth and VFA compared to their NWO counterparts. Among participants aged 2029 or 3039, obese participants had Nakanishi Y,

4) Lifestyle
Among participants aged 1819, lifestyle patterns did not differ significantly between NWO and non-obesity groups. Among those aged 2029, there was a significant intergroup difference in walking speed (p < 0.05): 42.9% of non-obese participants reported fast walking speeds, as compared to 21.2% of those with NWO. In the age group of 3039, there were significant intergroup differences in the two items. Only 1.7% of nonobese participants reported a gain of more than 10 kg since they were 20 years, compared with 18.4% of NWO participants (p < 0.001). Similarly, in the second item, 29% of non-obese participants reported a gain or loss of more than 3 kg in the past year, whereas 44.9% of their NWO counterparts reported the same (p < 0.05). Items that exhibited an intergroup difference were then subjected to a logistic regression. Among participants aged 2029, slow-walkers exhibited an odds ratio of 2.8 (p = 0.05; 95% CI: 1.07.5), denoting that those in this age group who walked slower than their peers were 2.8 times more likely to be NWO. Among those aged 3039, those who had gained 10 kg or more since the age of 20 were associated with a 13.0-fold increase in odds for NWO (p < 0.001; 95% CI: 3.4 50.1). Moreover, in this group, those who gained weight were 13 times more likely to be NWO than those who did not. Similarly, those who gained or lost over 3 kg in the past year exhibited an odds ratio of 2.0 (p = 0.037; 95% CI: 1.03.8), denoting that those who experienced weight fluctuation were twice as likely to be NWO (see Table-3).

NWO prevalence
This study was conducted on women under the age of 40. We found that 24.4% of them had NWO ─ defined as BMI of < 25 and a BFP of ≥30%. The NWO rate surpassed 30% in the 1819 and 2029 age groups, and came close to 20% in the 3039 age range.
As  7) . It is thus understood that NWO in Japan is higher than in Europe and the U.S. and similar to results in Korea.
Japanʼs Ministry of Health, Labor and Welfare conducts national health and nutrition surveys, providing benchmarks for adult health. The result of BMI and abdominal girth in this study are almost identical to the 2017 survey. Survey results revealed that adults under age 40 were less likely to exhibit abnormal laboratory readings, which indicates metabolic syndrome, compared to those over Nakanishi Y, et   Meanwhile, we found an intergroup difference in laboratory test results in the 3039 age range. NWO participants had significantly higher readings for triglycerides and LDL cholesterol, and significantly lower readings for HDL cholesterol, compared to their counterparts in the same age group. Regardless of age, NWO participants had significantly higher results for abdominal girth and VFA. Consistent with literature 8) , results showed that NWO was associated with increased risk of metabolic syndrome, suggesting that high triglycerides and LDL cholesterol and low HDL cholesterol were warning signs of NWO in women aged 3039.
Meanwhile, in comparison between NWO and obesity, regardless of age, obese participants had significantly higher abdominal girth and VFA compared to those with NWO. The intergroup difference was not significant for triglycerides and LDL cholesterol. As mentioned, women under 40 were far less likely to exhibit abnormal blood readings, which indicates metabolic syndrome, compared to those over 40 22) . Therefore, NWO may be mildly obese and have the risk of arteriosclerotic disease as with obesity.
Under Japanʼs health checkup program, a woman would qualify for stratified health guidance if her health checkup results indicated an abdominal girth larger than 90 cm and a BMI greater or equal to 25. In determining whether a woman qualifies for health guidance, the program considers abnormal laboratory readings and smoking as risk factors 23) . However, abnormal fat-related laboratory data alone are not sufficient to qualify. If women with such abnormalities miss out on health care guidance, they may struggle to improve their health. Therefore, criteria under which women under age 40 qualify for health guidance must be reconsidered.

Factors of NWO: Segmental LBM
For all body segments (upper limbs, lower limbs, trunk), NWO participants fell significantly lower of their benchmark segmental LBM compared to nonobese participants. Although BMI is a standard benchmark, the more useful physical indicators of NWO are high BFP and low LBM (an indicator of muscle mass), which makes it different from obesity, which involves excessive body weight 8) .
Results showed that among participants aged 1819, NWO was associated with significantly lower LBM in lower limbs, at the 0.1% level, and in the trunk, at the 1% level. Among those aged 2029, it was associated with significantly lower LBM in the trunk and lower limbs, both at the 0.1% level, and in the upper limbs, at the 5% level. Among those aged 3039, NWO participants had significantly lower LBM in all segments, at the 0.1% level. Thus, among NWO participants, age was associated with decreased LBM throughout the body; those aged 1819 had concerning levels of LBM in lower limbs; those aged 2029 had concerning levels of LBM in their lower limbs and trunk; and those aged 3039 had concerning levels in all segments. These findings suggest that body composition data, particularly segmental LBM, could predict NWO.
As aforementioned, NWO has been associated with unhealthy lifestyle patterns, such as insufficient exercise or aversion to exercise 9) 10) . Likewise, our results revealed walking speed to be a predictor of NWO among participants aged 2029. As NWO is associated with metabolic syndrome, women under the age of 40 require exercise guidance to help them maintain optimal segmental LBM. NWO participants, regardless of age, had particularly concerning LBM results for their lower limbs. Previous studies have highlighted that Japanese people experience aging-related decline in their muscle mass, particularly in the lower limbs 21) 24) . Changes in muscle mass with advancing age were different in each body part. In upper limbs, there was little change with advancing age in men and women. In lower limbs, decrease in muscle mass began after two decades, with greatest reduction in this muscle mass of all body parts with advancing age. In the trunk, the slope of the regression line increased from the second to the fifth decade, after which the slope decreased. Findings indicated that lower limb muscle mass was first to begin to decrease and also showed the greatest decrease 24) .
When this aging-related decline is considered, the findings imply that people should receive exercise guidance from age 10 onward to ensure they build enough muscle mass in their legs, thus preventing future metabolic syndrome onset.
In a recent study, Nakano et al. (2018) found that female college students gained muscle mass in lower limbs after undertaking a three-month exercise regimen consisting of body workouts, running, and walking 25) . In their study, increased muscle mass against body weight probably reflected significant decline in body weight. Given that lower limbs are relatively responsive to exercise instruction, it would probably be efficacious to focus lifestyle guidance on exercising lower limbs.

Factors of NWO: Sharp fluctuation in body weight in ages 3039
We found that NWO was associated with the experience of sharp weight fluctuation in ages 30 39 (gaining more than 10 kg since age 20 or experiencing weight fluctuation in the past year). Weight fluctuation in the past year might be a result of dieting. Studies on female college students found that NWO was associated with dieting and failed dieting 6) 8) 11) . Nishimura et al. (2010) reported that eating disorders increased odds of NWO onset 12) .
Our results indicated that NWO was associated with sharp increase in body weight during adulthood even if a personʼs weight was currently below average. Specifically, women aged 3039 who had gained 10 kg or more since age 20 were 13 times more likely to experience NWO. Consequently, health guidance should focus on instilling dietary and exercise behaviors that help individuals maintain the weight they had at age 20.

Limitations
One limitation is that because we collected data during health checkups, we relied on short-form questionnaire surveys to ascertain lifestyle patterns. Although we could identify the relationship between weight regulation and NWO, we could not obtain data on how dieting might lead to weight regulation failure. A future study should analyze NWO participantsʼ lifestyle patterns and dieting experience in greater detail to identify NWOʼs lifestyle factors. Another limitation is that the cross-sectional design employed cannot discuss causality, and was unable to precisely capture changes in individuals with NWO. Future studies should examine how such individuals change with time to identify how NWO in young people affects life course.

Conclusion
Focusing on women under 40 years and thus not yet subject to required health checkups, we examined NWO prevalence and NWO-related factors by age (1819, 2029, and 3039). The results are as follows. First, the prevalence of NWO in each age group ranged from 20% to 30% and was higher among participants aged 1819 or 2029 than those aged 3039. Second, regardless of age, NWO was associated with greater abdominal girth and VFA. Among participants aged 3039, and not among other age groups, NWO was associated with higher triglycerides and LDL cholesterol and lower HDL cholesterol. Third, NWO was associated with lower segmental LBM, particularly in lower limbs. Fourth, although lifestyle patterns were unrelated to NWO among participants aged 1819, they were related to NWO among those aged 2029 or 3039. In ages 3039, a gain of 10 kg since age 20 was associated with a 13-fold increase in the odds for NWO. These findings suggest that for women under 40 years, detecting NWO early and guidance on effective lifestyle and exercise are necessary.