Effective waist circumference reduction rate necessary to avoid the development of type 2 diabetes in Japanese men with abdominal obesity

Transcription

1 2017, 64 (9), Original Effective waist circumference reduction rate necessary to avoid the development of type 2 diabetes in Japanese men with abdominal obesity Risa Kashiwagi 1), Hiromi Iwahashi 2), Yuya Yamada 1), Takaaki Sakaue 1), Tomonori Okita 1), Yusuke Kawachi 1), Ryuya Iwamoto 1), Kenji Saisho 1), Sachiko Tamba 1), Koji Yamamoto 1), Takehiko Watanabe 3), Takashi Fujimoto 3) and Yuji Matsuzawa 1) 1) Department of Endocrinology and Metabolism, Sumitomo Hospital, Osaka, Japan 2) Department of Diabetes Care Medicine, Graduate School of Medicine, Osaka University, Suita, Japan 3) Physical Check up Center, Sumitomo Hospital, Osaka, Japan Abstract. The aim of this study was to determine the effective waist circumference (WC) reduction rate in avoiding the development of type 2 diabetes mellitus (T2DM) in <55 years and 55 years Japanese men with abdominal obesity. The study subjects were 795 men with WC 85 cm, fasting plasma glucose <126 mg/dl, 2-hr plasma glucose on 75 g of oral glucose tolerance test <200 mg/dl, and HbA1c % (38-40 mmol/mol) at baseline who underwent general health checkups more than twice between April 2007 and May They were divided into 5 groups based on the change in WC during the observation period (WC gain group, and four groups stratified according the rate of WC loss). The subjects were also divided into the <55 years and 55 years (at baseline) subgroups. The cumulative incidence rate of T2DM was analyzed and compared among the groups. The cumulative incidence rates of the largest WC loss quartile ( 5.45 %) in all age, of the largest WC loss quartile ( 5.60 %) and second largest WC loss quartile ( %) in the <55 years subgroup, and of the largest WC loss quartile ( 5.37 %) in the 55 years subgroup were significantly lower than that of the gain group (p<0.001, p=0.009, 0.012, and 0.012, respectively). WC reduction rate of at least about 3 % in the younger (<55 years) and at least about 5 % in the older ( 55 years) non-diabetic Japanese men with abdominal obesity can effectively reduce the chance of development of T2DM. Key words: Abdominal obesity, Type 2 diabetes mellitus, Waist circumference ABDOMINAL OBESITY, especially visceral fat accumulation, is involved in the pathophysiology of various health disorders, such as glucose and lipid dysmetabolism and cardiovascular diseases [1-7]. According to the recent studies, the prevalence of metabolic syndrome in Japanese type 2 diabetes patients is about 50% in men and 30% in women [8, 9]. Ministry of Health, Labour and Welfare started the health care program from 2008, which provides specific health checkups and guidance focusing on visceral fat obesity aged 40 or over, and it is known that these efforts have lead to the reduction of health expenditures [10]. Submitted Mar. 15, 2017; Accepted Jun. 2, 2017 as EJ Released online in J-STAGE as advance publication Jul. 15, 2017 Correspondence to: Risa Kashiwagi, M.D., Department of Endocrinology and Metabolism, Sumitomo Hospital, , Nakanoshima Kita-ku, Osaka-shi, Osaka, Japan. kashiwagi-risa@sumitomo-hp.or.jp The Japan Endocrine Society Previous studies indicated that reduction of body weight (BW) is important in preventing the development of type 2 diabetes mellitus (T2DM). In fact, our group has shown that BW reduction is closely associated with improvement of glycemic control in Japanese men with visceral fat accumulation and HbA1c of % (38-40 mmol/mol), but not in those without abdominal obesity [11]. We have also reported that at least 4-5 % BW loss within a period of 3 years is necessary to minimize the risk of T2DM in Japanese men with visceral fat accumulation [12]. These results suggest that lifestyle intervention for Japanese men with abdominal obesity is effective in preventing T2DM through the reduction of visceral fat. Providing the target value of waist circumference (WC) reduction in Japanese men with abdominal obesity is valid because WC is indispensable criteria to diagnose as metabolic syndrome, also in the process of specific health checkups [13].

2 882 Kashiwagi et al. However, the critical value of decrease in WC necessary to reduce the risk of T2DM remains unknown. Moreover, little is known whether such WC reduction is age-dependent. The concept of frailty and sarcopenia is being widely spread in clinical practice and research. Chronic health disorders including diabetes are associated with the age-dependent loss of muscle mass and function among older adults [14]. Thus, to determine the target WC reduction rate for different age groups is important when providing the specific health guidance. The present large-scale study of Japanese men with abdominal obesity was designed to determine the effective WC reduction rate necessary to significantly prevent the development of T2DM, in two different age groups. Materials and Methods Subjects The targeted subjects were 27,052 individuals (16,395 men and 10,657 women) who underwent health checkup at the Physical Check up Center, Sumitomo Hospital, between April 2007 and May The health checkup program for the Japanese citizens is designed to detect diseases or risk factors at early stages and they often take the checkups spontaneously and none of these participants had any serious diseases. The eligibility criteria for enrolment in the present study included males who had taken at least two checkup examinations during the above period, with WC of 85 cm, fasting plasma glucose (FPG) <126 mg/dl, HbA1c % (38-40 mmol/mol) and 2-hr plasma glucose (PG) on 75 g of oral glucose tolerance test (OGTT) of <200 mg/dl, if available, at baseline. We excluded women subjects, because the number of men who underwent our health checkup is larger than that of women. FPG 126 mg/dl or 2-hr PG on OGTT 200 mg/dl was used for the diagnosis of T2DM according to the criteria for the diagnosis of diabetes in Japan. According to the report of Japan Diabetes Society (JDS), individuals with HbA1c % should be advised to receive health guidance to prevent T2DM. Subjects using antidiabetic medications or having regular visits to the hospital or family physician for T2DM were excluded from the study. We also excluded participants with HbA1c of 8 % (64 mmol/mol), at the onset of T2DM, because we considered that reduction of WC or BW could results from deterioration of glucose tolerance rather than lifestyle modification. Using these criteria, 795 participants were included in the present study. The study was approved by the human ethics committees of Sumitomo Hospital and Osaka University and was conducted according to the principles of the Declaration of Helsinki. Informed consent to provide medical information and blood samples was obtained before the checkup examinations from all participants and each subject had the right to refuse the use of their results. Clinical and laboratory assessment Height, BW, WC, and blood pressure were measured after overnight fasting. WC was measured at the level of the umbilicus based on the recommendation of the Japan Society for the Study of Obesity [15]. The IDF/AHA Harmonized Criteria recommend setting the cutoff value of WC for obesity by using population- and country-specific definitions [16]. In Japan, the cutoff values of WC for obesity are 85 cm for men and 90 cm for women. These cutoff values correspond to 100 cm 2 of visceral fat area at the umbilicus level measured by computed tomography (CT) [17]. In Japan, 100 cm 2 of visceral fat area is associated with the best sensitivity and specificity for obesity-related multiple risk factors [15]. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. We also used questionnaires to obtain information about use of medications, family history of T2DM, and regular visits to the hospital or family physician. Blood samples were taken and used for measurement of fasting plasma glucose (FPG), fasting serum immunoreactive insulin (F-IRI), HbA1c, lipid profile, urine acid (UA), and creatinine (Cr). The value for HbA1c (%) represented the National Glycohemoglobin Standardization Program (NGSP). The formula used for conversion of HbA1c (JDS) to HbA1c (NGSP) was as follows: HbA1c (NGSP) (%) = 1.02 HbA1c (JDS) (%) % [18, 19]. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as F-IRI (µu/ ml) FPG (mg/dl) / 405. Serum total adiponectin levels were measured by the latex particle-enhanced turbidimetric immunoassay (human adiponectin latex kit; Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan). The within-run and total coefficient of variation of this measurement were % and %, respectively. The results correlated highly with those obtained by enzyme-linked immunosorbent assay (r=0.99) [20].

3 Waist reduction and type 2 diabetes 883 Endpoints and categorization according to WC change and age Based on the clinical data, the participants were at high risk for the development of T2DM in association with visceral fat accumulation. We advised these individuals to improve their dietary habits and engage in adequate aerobic exercise in order to avoid the potential development of T2DM. Each participant was followed up from the first visit to Physical Check up Center in our hospital and observed until the diagnosis of T2DM or the last visit for checkup by May In this study, diabetes was set at HbA1c 6.5 %, FPG 126 mg/dl, or 2-hr PG on OGTT 200 mg/dl if conducted. The mean follow-up period was 3.6 ± 2.3 (min-max: ) years. The percent change in WC (%ΔWC) was estimated at the end of the observation period. %ΔWC was calculated as follows: [(WC at endpoint - WC at baseline)] / WC at baseline. We divided the participants into five groups according to changes in WC during the observation period; the WC gain group and the quartile of WC loss groups. We also divided the participants into two subgroups by age; <55 years (the younger age subgroup) and 55 years (the older age subgroup) at baseline. Statistical analysis Data are shown as mean ± standard deviation, and compared by the Student s t-test. We compared the cumulative incidence rates of T2DM between participants who lost WC and those who gained using Kaplan-Meier product limit method and log-rank test. The cumulative incidence rate for each WC loss quartile was compared with that of the WC gain group. We used Cox proportional hazard analysis to estimate the multivariable adjusted hazard ratio (HR) and 95 % confidence interval (CI). The cumulative incidence rate according to change in WC was analyzed in each age subgroup in the same way as mentioned above. The factors associated with the development of T2DM were evaluated using multivariate Cox proportional hazards regression models. Multivariable models included the covariates of baseline age, FPG, HbA1c, F-IRI, serum adiponectin, systolic blood pressure (SBP), total cholesterol, triglyceride (TG), high-density lipoprotein (HDL) cholesterol, Cr, WC, family history of T2DM, and %ΔWC. We then used the stepwise backward selection procedure to identify the variables that independently and significantly influenced the development of T2DM based on the p-value. Subsequently, we performed the same methods mentioned above to determine the impact of BW change. We performed Kruskal-Wallis test and confirmed that there were no significant differences in the time periods from the previous visits to endpoints between each group. All statistical analyses were performed using SPSS software version 23 for Windows (SPSS Inc., Chicago, IL). Significance level was set at p-value <0.05. Results Cumulative incidence rate of T2DM according to WC changes The baseline clinical characteristics are summarized in Table 1. There were no significant differences in age, BMI, WC, FPG, HbA1c, SBP, UA, and lipid profile between the WC loss and gain groups. Changes in BW were significantly associated with changes in WC. There was no significant difference in the proportion of participants with family history of T2DM between the WC loss and gain group. There were no significant differences in BMI, FPG, HbA1c, SBP, and lipid profile among the four WC loss groups, compared with the WC gain group. Changes in BW were significantly associated with changes in WC in each quartile group. There were no significant differences in the proportion of individuals with family history of T2DM between each quartile group and the gain group. Fig. 1A shows the cumulative incidence rate and proportion of diabetes-free subjects of each quartile group and the WC gain group. The cumulative incidence rate was significantly lower in the largest WC loss quartile (5.45 % and over) than in the gain group (p<0.001). There was no significant difference in the cumulative incidence rate between the other WC loss quartiles [second largest (3.18 %-5.38 %), third largest ( %), smallest ( %)] and the gain group. Table 2 shows the results of multivariate Cox proportional hazards regression analyses. Baseline FPG, HbA1c and %ΔWC were identified as significant factors in the development of T2DM. The results of stepwise backward selection procedure are also shown in Table 2. Baseline FPG, HbA1c, serum adiponectin level and %ΔWC were the most significant factors.

4 884 Kashiwagi et al. Table 1 Baseline clinical characteristics Total WC gain group All WC loss groups 1Q 5.45 % 2Q % 3Q % 4Q % n Time from last visits to endpoint (years) 1.5 ± ± ± ± ± ± ± 1.0 Age (years) 56 ± ± ± ± ± ± 10* 57 ± 9 BMI (kg/m 2 ) 25.5 ± ± ± ± ± ± ± 2.3 WC (cm) 92.0 ± ± ± ± ± ± 5.3* 91.6 ± 4.7 SBP (mmhg) 129 ± ± ± ± ± ± ± 13 DBP (mmhg) 81 ± 9 81 ± ± 9 82 ± ± 9 82 ± 9 80 ± 8 Total-cholesterol (mg/dl) 217 ± ± ± ± ± ± ± 35 HDL-cholesterol (mg/dl) 55 ± ± ± ± ± ± ± 13 LDL-cholesterol (mg/dl) 136 ± ± ± ± ± ± ± 33 TG (mg/dl) 158 ± ± ± ± ± ± ± 84 UA (mg/dl) 6.5 ± ± ± ± ± 1.1* 6.4 ± ± 1.1 FPG (mg/dl) 102 ± ± ± ± ± ± ± 10 HbA1c (%) 5.8 ± ± ± ± ± ± ± 0.2 Cr (mg/dl) 0.87 ± ± ± 0.16* 0.83 ± ± ± ± 0.13 HOMA-IR 2.0 ± ± ± ± ± ± ± 1.8 F-IRI (μu/ml) 8.0 ± ± ± ± ± ± ± 6.0 Adiponectin (μg/ml) 6.9 ± ± ± ± ± ± ± 2.7 %ΔWC -1.2 ± ± ± 3.1* -8.3 ± 2.8* -4.1 ± 0.7* -2.6 ± 0.4* -1.1 ± 0.5* %ΔBW -1.1 ± ± ± 4.3* -6.6 ± 5.1* -3.3 ± 3.4* -1.3 ± 3.1* -0.8 ± 2.5* Family history of T2DM (n) Data are mean ± standard deviation or number of participants. * p<0.05, compared with the WC gain group. BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglyceride; UA, urine acid; FPG, fasting plasma glucose; Cr, creatinine; F-IRI, fasting serum immunoreactive insulin; HOMA-IR, homeostasis model assessment of insulin resistance; %ΔBW, percent change in body weight; %ΔWC, percent change in waist circumference; T2DM, type 2 diabetes mellitus.

5 Waist reduction and type 2 diabetes 885 Fig. 1 Kaplan-Meier curves of the proportion of diabetes-free participants of the WC loss quartiles and WC gain group A, Analysis of data of all participants: Cumulative incidence rates of the WC loss quartiles [largest (5.45 % and over), second ( %), third ( %), smallest ( %)], and WC gain group were, 7.0 %, 13.8 %, 16.4 %, 23.3 %, and 19.8 %, respectively. The cumulative incidence rate was significantly lower in the largest WC loss quartile (5.45 % and over) than in gain group (p<0.001). B, Analysis of data of the younger age subgroup (<55 years): The cumulative incidence rates of the WC loss quartiles [largest (5.60 % and over), second ( %), third ( %), smallest ( %)], and WC gain group were, 6.2 %, 8.3 %, 16.3 %, 12.2 %, and 20.3 %, respectively. The cumulative incidence rate was significantly lower in the largest WC loss quartile (5.60 % and over) than that of the gain group (p=0.009), and the cumulative incidence rate of the second largest WC loss quartile ( %) was also significantly lower than that of the gain group (p=0.012). C, Analysis of data of the older age subgroup ( 55 years): The cumulative incidence rates of the WC loss quartiles [largest (5.37 % and over), second ( %), third ( %), smallest ( %)], and the WC gain group were, 7.7 %, 18.2 %, 16.2 %, 30.4 %, and 19.5 %, respectively. The cumulative incidence rate was significantly lower in the largest WC loss quartile (5.37 % and over) than in gain group (p=0.012). WC, waist circumference.

6 886 Kashiwagi et al. Table 2 Hazard ratios for the development of type 2 diabetes mellitus Hazard ratio 95 % Confidence interval p-value Age NS FPG <0.001 HbA1c F-IRI NS Adiponectin NS SBP NS Total-cholesterol NS TG NS HDL-cholesterol NS Cr NS WC NS Family history of T2DM NS %ΔWC <0.001 Stepwise backward selection procedure Adiponectin FPG <0.001 HbA1c %ΔWC <0.001 FPG, fasting plasma glucose; F-IRI, fasting serum immunoreactive insulin; SBP, systolic blood pressure; TG, triglyceride; HDL, high-density lipoprotein; Cr, creatinine; WC, waist circumference; T2DM, type 2 diabetes mellitus; %ΔWC, percent change in waist circumference; NS, not significant. Cumulative incidence rate of T2DM based on age Table 3 shows the baseline clinical characteristics of the younger (<55 years) and the older ( 55 years) subjects. There were no significant differences in age, WC, FPG, HbA1c, SBP, and UA between the WC loss and gain groups in both age subgroups. For the older age subgroup, subjects of the WC loss group were significantly younger with lower BMI at baseline compared with those of the gain group (p=0.033 and p=0.006, respectively). In both age subgroups, changes in BW were significantly associated with changes in WC. For the younger age subgroup, the proportion of subjects with family history of T2DM was not different between the WC loss and gain groups; however, the proportion was significantly higher in the WC loss group in the older age subgroup (p=0.019). Table 3 also shows the baseline clinical characteristics of subjects of the WC loss quartile groups stratified by the two age subgroups. For each quartile, there were no significant differences in FPG, HbA1c, and SBP, compared with the WC gain group, and changes in BW were significantly associated with WC change in each quartile in both age subgroups. For the older age subgroup ( 55 years), we found significant differences in BMI, UA, and total cholesterol between the quartile groups and WC gain group. Also in the older age subgroup, there were significant differences in the proportion of family history between the third largest WC loss quartile ( %) and the smallest WC loss quartile ( %), and the gain group (p=0.031 and 0.004, respectively), but not for the younger age subgroup. Fig. 1B shows the cumulative incidence rate and proportion of diabetes-free subjects of each quartile group and the WC gain group for the younger age subgroup (<55 years). The cumulative incidence rate was significantly lower in the largest WC loss quartile (5.60 % and over) than the gain group (p=0.009), and the rate of the second largest WC loss quartile ( %) was also significantly lower than that of the gain group (p=0.012). Fig. 1C shows the cumulative incidence rate and proportion of diabetes-free subjects of each quartile group and WC gain group for the older age subgroup ( 55 years). The cumulative incidence rate was significantly lower only in the largest WC loss quartile (5.37 % and over) than in the gain group (p=0.012). Since baseline age was significantly lower in the older WC loss group, we also analyzed their data after excluding three oldest subjects to adjust age. The baseline clinical characteristics, cumulative incidence rate, and proportion of diabetes-free participants were similar to those of the entire group. We performed the same analyses replacing WC change with BW change, and similar results were obtained. The results are shown in Supplementary Tables 1, 2 and Supplementary Fig. 1. Analysis of data of all ages showed a significantly lower cumulative incidence rate for the largest BW loss quartile (5.31 % and over) compared with the BW gain group

7 Waist reduction and type 2 diabetes 887 Table 3 Baseline clinical characteristics of the younger (<55 years) and the older ( 55 years) age subgroups Total WC gain group All WC loss groups Largest WC loss quartile Second largest WC loss quartile Third Smallest WC largest WC loss quartile loss quartile The younger age subgroup (<55 years) Range of WC loss quartile (%) n Time from last visit to endpoint (years) 1.6 ± ± ± ± ± ± ± 1.4 Age (years) 47 ± 6 47 ± 6 47 ± 6 46 ± 7 47 ± 6 49 ± 5 48 ± 5 BMI (kg/m 2 ) 26.2 ± ± ± ± ± ± ± 2.8 WC (cm) 92.8 ± ± ± ± ± ± 6.7* 93.3 ± 5.5 SBP (mmhg) 129 ± ± ± ± ± ± ± 15 DBP (mmhg) 82 ± ± ± ± ± ± ± 10 Total-cholesterol (mg/dl) 222 ± ± ± ± ± ± ± 36 HDL-cholesterol (mg/dl) 54 ± ± ± ± ± ± ± 13 LDL-cholesterol (mg/dl) 142 ± ± ± ± ± ± ± 36 TG (mg/dl) 177 ± ± ± ± ± ± ± 87 UA (mg/dl) 6.7 ± ± ± ± ± ± ± 1.0 FPG (mg/dl) 101 ± ± ± ± ± ± ± 10 HbA1c (%) 5.8 ± ± ± ± ± ± ± 0.2 Cr (mg/dl) 0.85 ± ± ± ± ± ± ± 0.12 HOMA-IR 2.3 ± ± ± ± ± ± ± 2.5 F-IRI (μu/ml) 8.9 ± ± ± ± ± ± ± 8.1 Adiponectin (μg/ml) 6.3 ± ± ± ± ± ± ± 2.2 %ΔWC -1.3 ± ± ± 3.0* -8.5 ± 2.3* -4.3 ± 0.6* -2.8 ± 0.4* -1.2 ± 0.5* %ΔBW -1.2 ± ± ± 4.4* -7.7 ± 4.4* -3.5 ± 3.5* -1.9 ± 3.4* -0.7 ± 2.8* Family history of T2DM (n) The older age subgroup ( 55 years) Range of WC loss quartile (%) n Time from last visit to endpoint (years) 1.4 ± ± ± ± ± ± ± 0.6 Age (years) 63 ± 6 62 ± 6 63 ± 7* 63 ± 7 63 ± 6 64 ± 8 63 ± 6 BMI (kg/m 2 ) 25.0 ± ± ± 2.0* 24.9 ± ± ± ± 1.6* WC (cm) 91.5 ± ± ± ± ± ± ± 3.8 SBP (mmhg) 128 ± ± ± ± ± ± ± 13 DBP (mmhg) 81 ± 8 81 ± 9 80 ± 8 81 ± 8 81 ± 9 80 ± 8 79 ± 7 Total-cholesterol (mg/dl) 214 ± ± ± ± ± 33* 215 ± ± 33 HDL-cholesterol (mg/dl) 56 ± ± ± ± ± ± ± 13 LDL-cholesterol (mg/dl) 132 ± ± ± ± ± ± ± 30 TG (mg/dl) 144 ± ± ± ± ± ± ± 79 UA (mg/dl) 6.4 ± ± ± ± ± 1.1* 6.3 ± ± 1.1 FPG (mg/dl) 102 ± ± ± ± ± ± ± 10 HbA1c (%) 5.8 ± ± ± ± ± ± ± 0.2 Cr (mg/dl) 0.88 ± ± ± ± ± ± ± 0.14 HOMA-IR 1.9 ± ± ± ± ± ± ± 1.0 F-IRI (μu/ml) 7.3 ± ± ± ± ± ± ± 3.8 Adiponectin (μg/ml) 7.4 ± ± ± ± ± ± ± 2.9 %ΔWC -1.1 ± ± ± 3.1* -8.2 ± 3.2* -4.0 ± 0.7* -2.6 ± 0.4* -1.0 ± 0.5* %ΔBW -1.1 ± ± ± 4.2* -5.8 ± 5.6* -3.0 ± 3.3* -1.4 ± 3.1* -0.8 ± 2.3* Family history of T2DM (n) * * 20* Data are mean ± standard deviation or number of participants. * p<0.05, compared with the WC gain group. BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglyceride; UA, urine acid; FPG, fasting plasma glucose; Cr, creatinine; F-IRI, fasting serum immunoreactive insulin; HOMA-IR, homeostasis model assessment of insulin resistance; %ΔBW, percent change in body weight; %ΔWC, percent change in waist circumference; T2DM, type 2 diabetes mellitus.

8 888 Kashiwagi et al. (p<0.001). Moreover, the cumulative incidence rate was significantly lower in the second largest BW loss quartile ( %) (p=0.001). Analysis of the younger age subgroup (<55 years) also showed significant differences in the cumulative incidence rate between the largest BW loss quartile (6.31 % and over) and gain group (p<0.001). Furthermore, the cumulative incidence rate was significantly lower in the second largest BW loss quartile ( %) (p=0.003). Analysis of the older age subgroup ( 55 years) showed significantly lower cumulative incidence rate only in the largest BW loss quartile (4.72 % and over) (p<0.001). Discussion The adipose tissue is not only a storage organ of excessive energy but also an endocrinological organ secreting many active cytokines. Accumulation of visceral fat leads to the development of T2DM, dyslipidemia, hypertension, and arteriosclerosis. Therefore, reduction of visceral fat is expected to improve or prevent these disorders. In Japan, statistical surveys conducted by Ministry of Health, Labour and Welfare found that the next year health expenditures of people who were provided the specific health checkups and specific health guidance focused on metabolic syndrome were about 30% smaller than those of people who were not [10]. Thus, to decide the target of WC reduction rate is important for health providers. Our study showed that reduction of WC of at least about 3 % in the younger (<55 years) men and at least about 5 % in the older ( 55 years) non-diabetic Japanese men with abdominal obesity could reduce the chance of development of T2DM. These target levels of WC reduction would be useful for specific health guidance focused on metabolic syndrome. This study has several important features that need to be mentioned. First, to our knowledge, this is the first prospective study that examined the effect of WC reduction as well as BW reduction on the development of T2DM in different age subgroups. The prevalence of T2DM was reported to be high in subjects with visceral fat area of 100 cm 2 [21, 22]. Data of the Epidemiological Study on the Insulin Resistance Syndrome (DESIR) showed that increase in WC was significantly associated with 9-year T2DM incidence, with a standardized odds ratio of 1.79 ( ) [21]. Several studies also reported that decrease in visceral fat improves glucose tolerance [23, 24]. The Diabetes Prevention Program reported that BW loss through lifestyle intervention was highly effective in preventing T2DM, especially in patients with impaired glucose tolerance [25]. However, changes in BW do not necessarily correspond to changes in visceral fat, because BW includes other constituents such as fluid, muscle, and subcutaneous fat in addition to visceral fat. Previous study has reported that ectopic lipid deposit in skeletal muscle and liver is accompanied by visceral fat accumulation, and there is a strong relationship between liver lipid deposition and the insulin resistance [26]. It has been also reported that lipid deposition in muscle increases with age [26]. Moreover, frail status, which is associated with age-related decline in skeletal muscle mass, can also influence BW change in the elderly [27]. In addition, it has been reported that WC correlates with visceral fat in Japanese men [28], suggesting that reduction in WC seems to reflect decrease in visceral fat. In this study, similar results were obtained when we performed the same analyses replacing WC change with BW change. However, we found that there were 26% patcipants whose WC change and BW change did not correspond (99 men whose WC gained in spite of BW reduction and 108 men whose WC reduced in spite of BW gain). From the perspective of providing specific health checkups and guidance focusing on visceral fat obesity, we think that it is reasonable to focus on WC change instead of BW change. Moreover, WC is a theoretically better marker because WC gain with BW reduction may reflect frail status as mentioned above. Nevertheless, the extent of WC reduction necessary to reduce the likelihood of development of T2DM has not been proposed yet. CT at the umbilical level is the standard method for evaluating the amount of visceral fat [29]. However, it has problems of high costs and radiation exposure. Several reports used instead the bioelectrical impedance method (BIA) [30, 31], but this method is not widely used at present. On the other hand, measurement of WC is a simple and easy way for both the health care provider and patient to estimate visceral fat accumulation. Our results indicate that at least about 3 % WC reduction in the younger (<55 years) men, and at least about 5 % in the older ( 55 years) men would be clinically suitable target for non-diabetic Japanese men with abdominal obesity. Both in the younger (<55 years) and in

9 Waist reduction and type 2 diabetes 889 the older ( 55 years) men, the mean WC at endpoint of the largest WC loss quartile is lower than 85cm by calculating mean baseline WC and mean %ΔWC (84.3 cm and 84.8 cm, respectively). These results indicate that reduction of WC below the cutoff value that corresponds to 100 cm 2 of visceral fat area may relate to the reduction of the risk of T2DM. However, we found no significant difference in cumulative incident rate of T2DM whether WC of the participants achieved less than 85 cm or not, although the former had a tendency of low incidence rate of T2DM. We think these facts suggest that WC reduction of 5.45% is effective for preventing T2DM even in individuals whose WC did not achieve less than 85 cm. In addition, we also think that achievement of WC less than 85 cm at endpoint may be another important target for decreasing multiple risk factors, because it is known that the cutoff value of WC corresponds to 100 cm 2 of visceral fat area, the individuals with which proved to have one risk factor. Second, we found that the target percent reduction in WC was different among the different age groups. We have also reported that WC reduction of 5.5% and over reduced the risk of T2DM in three years [12], and it corresponds to the result of this study that WC reduction of 5.45 % and over could reduce the risk of T2DM in all ages. However, whether WC reduction necessary to reduce the risk of T2DM is age-dependent or not was unclear. In general, the results indicated that at least about 3 % loss in WC in the younger (<55 years) men, and at least about 5 % loss in WC in the older ( 55 years) men should be the target to prevent the development of T2DM. These reductions in WC were associated with a decrease in the cumulative incidence rate of about 60 %. Our study showed that more WC reduction is required in the older age subgroup compared with the younger age subgroup to avoid T2DM. In the elderly, glucose tolerance often deteriorates due to impairment in insulin action, secretion, and clearance [32]. Another study showed increased insulin resistance in Japanese aged 65 years, irrespective of abdominal obesity [33]. It is noteworthy that the older men need larger reduction in WC to avoid T2DM by correcting other factors in addition to the reduction of visceral fat accumulation. With regard to WC reduction, physicians should be careful to avoid WC loss programs/treatments that can cause frailty in this age population. This is important since frailty is highly prevalent in the elderly and a risk factor for adverse health outcomes, such as acute and chronic diseases, disability and mortality [34, 35]. In this regard, it has been suggested that regular in exercise has potential benefits in preventing frailty [36]. Thus, advice on not only dietary habits but also exercise activities are more likely to result in desirable WC loss in the older men. Moreover, the result of stepwise backward selection procedure identified baseline serum adiponectin concentration as a significant factor for the onset of T2DM, suggesting that serum adiponectin level can be potentially used to predict possible development of T2DM. Adiponectin is a multifunctional adipose-derived protein involved in prevention of atherogenesis, insulin-sensitization, enhancement of lipid-oxidation, and vasodilatory activities [37]. Since there is a strong relationship between adiponectin and visceral fat accumulation, we need further study to find out whether adiponectin can predict the development of T2DM. In this study, several baseline clinical parameters of subjects of the WC loss group and quartile subgroups were significantly different compared to those of subjects of the WC gain group. However, there were no specific increase or decrease tendency of those parameters in each quartile. We also analyzed the data after adjustment for different parameters at baseline, and similar results of cumulative incidence rate and proportion of diabetes-free participants were obtained. Although there were significant differences in the proportion of the older ( 55 years) subjects with family history between the WC loss quartiles and gain group, the results of multivariate Cox proportional hazards regression models indicated that family history is not a prognostic factor and therefore we think that it is not necessary to take such difference into consideration. This study has several limitations. First, we included subjects with FPG of <126 mg/dl, HbA1c % and 2-hr PG on 75 g OGTT <200 mg/dl, when measured, at baseline and did not perform OGTT in all the subjects. Therefore, not all subjects can be regarded as prediabetics. Second, the inclusion of subjects who underwent health checkups may imply selection bias toward highly health-conscious people. The extent of WC reduction that can prevent T2DM might be different in other cohorts, but we think that our population is appropriate to investigate the relationship between WC reduction and the development of T2DM.

10 890 Kashiwagi et al. In conclusion, WC reduction of at least about 3 % in the younger (<55 years) men and about 5 % in the older ( 55 years) non-diabetic Japanese men with abdominal obesity can reduce the chance of development of T2DM. These findings suggest that the extent of WC reduction necessary for minimization of diabetes risk is different among the different age groups. The above target levels of WC reduction are potentially beneficial for the prevention of T2DM in men and useful in providing specific health guidance. Acknowledgments We thank Prof. Ayumi Shintani, Department of Clinical Epidemiology and Biostatistics, Graduate School of Medicine, Osaka University, Suita, Japan for advice on statistical analysis. Disclosure The authors declare no relevant conflict of interest to disclose. Supplementary Table 1 Baseline clinical characteristics of subjects of the BW loss and gain groups Total BW gain group BW loss group Total 1Q 5.31 % 2Q % 3Q % 4Q % n Time from last visit to endpoint (years) 1.5 ± ± ± ± ± ± ± 1.2 Age (years) 56 ± ± 9 57 ± 10* 55 ± ± ± 10* 56 ± 9 BMI (kg/m 2 ) 25.5 ± ± ± ± 3.1* 25.5 ± ± ± 3.1 WC (cm) 92.0 ± ± ± ± 6.9* 92.1 ± ± ± 6.9 SBP (mmhg) 129 ± ± ± ± 15* 130 ± ± ± 15 DBP (mmhg) 81 ± 9 81 ± ± 9 83 ± 10* 82 ± 9 80 ± 8 80 ± 9 Total-cholesterol (mg/dl) 217 ± ± ± ± ± ± ± 34 HDL-cholesterol (mg/dl) 55 ± ± ± ± ± ± ± 11 LDL-cholesterol (mg/dl) 136 ± ± ± ± ± ± ± 32 TG (mg/dl) 158 ± ± ± ± ± ± ± 128 UA (mg/dl) 6.5 ± ± ± ± ± ± ± 1.1 FPG (mg/dl) 102 ± ± ± ± ± ± ± 10 HbA1c (%) 5.8 ± ± ± ± ± ± 0.2* 5.8 ± 0.2 Cr (mg/dl) 0.87 ± ± ± ± ± ± ± 0.13 HOMA-IR 2.0 ± ± ± ± ± ± ± 1.4 F-IRI (μu/ml) 8.0 ± ± ± ± ± ± 3.4* 7.9 ± 5.3 Adiponectin (μg/ml) 6.9 ± ± ± ± ± ± ± 3.2 %ΔWC -1.2 ± ± ± 4.1* -6.4 ± 4.3* -3.6 ± 3.0* -1.0 ± 3.2* -1.4 ± 3.2* %ΔBW -1.1 ± ± ± 3.6* -9.0 ± 3.5* -4.0 ± 0.6* -2.2 ± 0.5* -0.6 ± 0.4* Family history of T2DM (n) Data are mean ± standard deviation or number of subjects. * p<0.05, compared with BW gain group. BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglyceride; UA, urine acid; FPG, fasting plasma glucose; Cr, creatinine; F-IRI, fasting serum immunoreactive insulin; HOMA-IR, homeostasis model assessment of insulin resistance; %ΔBW, percent change in body weight; %ΔWC, percent change in waist circumference; T2DM, type 2 diabetes mellitus.

11 Waist reduction and type 2 diabetes 891 Supplementary Table 2 Comparison of baseline characteristics of the younger (<55 years) and the older ( 55 years) age subgroups Total BW gain group All BW loss groups Largest BW loss quartile Second largest BW loss quartile Third largest BW loss quartile Smallest BW loss quartile The younger age subgroup (<55 years) Range of WC loss quartile (%) n Time from last visit to endpoint (years) 1.7 ± ± ± ± ± ± ± 1.5 Age (years) 47 ± 6 47 ± 6 47 ± 6 47 ± 6 46 ± 7 48 ± 5 47 ± 5 BMI (kg/m²) 26.2 ± ± ± 3.4* 26.8 ± ± ± ± 3.9 WC (cm) 92.8 ± ± ± ± ± ± ± 9.0 SBP (mmhg) 129 ± ± ± ± 16* 129 ± ± ± 15 DBP (mmhg) 82 ± ± ± ± 11* 82 ± ± 8 79 ± 10 Total-cholesterol (mg/dl) 222 ± ± ± ± ± ± ± 37 HDL-cholesterol (mg/dl) 54 ± ± ± ± 8* 57 ± ± ± 11 LDL-cholesterol (mg/dl) 142 ± ± ± ± ± ± ± 36 TG (mg/dl) 177 ± ± ± ± ± 66* 169 ± ± 185 UA (mg/dl) 6.7 ± ± ± ± ± ± ± 1.1 FPG (mg/dl) 101 ± ± ± ± ± ± ± 10 HbA1c (%) 5.8 ± ± ± ± ± ± ± 0.2 Cr (mg/dl) 0.85 ± ± ± ± ± ± ± 0.15 HOMA-IR 2.3 ± ± ± ± ± ± ± 1.7 F-IRI (μu/ml) 8.9 ± ± ± ± ± ± ± 6.5 Adiponectin (μg/ml) 6.3 ± ± ± ± ± 2.2* 6.1 ± ± 2.9 %ΔWC -1.3 ± ± ± 3.9* -7.1 ± 3.5* -3.8 ± 3.1* -1.5 ± 3.2* -1.3 ± 2.7* %ΔBW -1.2 ± ± ± 3.7* -9.5 ± 2.8* -4.6 ± 0.9* -2.4 ± 0.7* -0.6 ± 0.4* Family history of T2DM (n) The older age subgroup ( 55 years) Range of WC loss quartile (%) n Time from last visit to endpoint (years) 1.4 ± ± ± ± ± ± ± 0.8 Age (years) 63 ± 6 62 ± 6 63 ± 7* 63 ± 6* 64 ± 7* 64 ± 7* 62 ± 6 BMI (kg/m²) 25.0 ± ± ± ± ± ± ± 2.3 WC (cm) 91.5 ± ± ± ± 5.5* 91.1 ± ± ± 4.8 SBP (mmhg) 128 ± ± ± ± ± ± ± 15 DBP (mmhg) 81 ± 8 80 ± 9 81 ± 8 81 ± 8 82 ± 9 80 ± 7 81 ± 8 Total-cholesterol (mg/dl) 214 ± ± ± ± ± ± ± 29* HDL-cholesterol (mg/dl) 56 ± ± ± ± ± ± ± 12 LDL-cholesterol (mg/dl) 132 ± ± ± ± ± ± ± 26 TG (mg/dl) 144 ± ± ± ± ± ± ± 65 UA (mg/dl) 6.4 ± ± ± ± ± ± ± 1.0 FPG (mg/dl) 102 ± ± ± ± ± ± ± 9 HbA1c (%) 5.8 ± ± ± ± ± ± ± 0.2 Cr (mg/dl) 0.88 ± ± ± ± ± ± ± 0.12 HOMA-IR 1.9 ± ± ± ± ± ± ± 1.1 F-IRI (μu/ml) 7.3 ± ± ± ± ± ± ± 4.0 Adiponectin (μg/ml) 7.4 ± ± ± ± ± 4.3* 7.7 ± ± 3.2 %ΔWC -1.1 ± ± ± 4.2* -5.9 ± 4.8* -3.2 ± 3.0* -0.9 ± 3.2* -1.2 ± 3.4* %ΔBW -1.1 ± ± ± 3.5* -8.4 ± 4.0* -3.7 ± 0.6* -2.0 ± 0.4* -0.6 ± 0.3* Family history of T2DM (n) * Data are mean ± standard deviation or number of participants. * p<0.05, compared with the BW gain group. BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglyceride; UA, urine acid; FPG, fasting plasma glucose; Cr, creatinine; F-IRI, fasting serum immunoreactive insulin; HOMA-IR, homeostasis model assessment of insulin resistance; %ΔBW, percent change in body weight; %ΔWC, percent change in waist circumference; T2DM, type 2 diabetes mellitus.

12 892 Kashiwagi et al. Supplementary Fig. 1 Kaplan-Meier curves of the proportion of diabetes-free participants of the BW loss quartiles and BW gain group A, Analysis of data of all participants: The cumulative incidence rates of the BW loss quartiles [largest (5.31 % and over), second ( %), third ( %), smallest ( %)], and BW gain group were, 4.5 %, 11.5 %, 18.4 %, 13.9 %, and 23.7 %, respectively. The cumulative incidence rate was significantly lower in the largest BW loss quartile (5.31 % and over) than the gain group (p<0.001). The cumulative incidence rate of the second largest BW loss quartile ( %) was also significantly lower than that of the gain group (p=0.001). B, Analysis of data of the younger age subgroup (<55 years): The cumulative incidence rates of the BW loss quartiles [largest (6.31 % and over), second ( %), third ( %), smallest ( %)], and BW gain group were, 4.3 %, 8.5 %, 10.6 %, 6.5 %, and 24.0 %, respectively. The cumulative incidence rate was significantly lower in the largest BW loss quartile (6.31 % and over) than the gain group (p<0.001). The cumulative incidence rate of the second largest BW loss quartile ( %) was also significantly lower than that of the gain group (p=0.003). C, Analysis of data of the older age subgroup ( 55 years): The cumulative incidence rates of the BW loss quartiles [largest (4.72 % and over), second ( %), third ( %), smallest ( %)], and BW gain group were, 7.6 %, 13.4 %, 22.4 %, 18.2 %, and 23.4 %, respectively. The cumulative incidence rate was significantly lower in the largest BW loss quartile (4.72 % and over) than the gain group (p<0.001). BW, body weight.

13 Waist reduction and type 2 diabetes 893 References 1. Nakamura T, Tokunaga K, Shimomura I, Nishida M, Yoshida S, et al. (1994) Contribution of visceral fat accumulation to the development of coronary artery disease in non-obese men. Atherosclerosis 107: Matsuzawa Y, Nakamura T, Shimomura I, Kotani K (1995) Visceral fat accumulation and cardiovascular disease. Obes Res 3 Suppl 5: 645S-647S. 3. Pascot A, Després JP, Lemieux I, Bergeron J, Nadeau A, et al. (2000) Contribution visceral obesity to the deterioration of the metabolic risk profile in men with impaired glucose tolerance. Diabetologia 43: Pascot A, Després JP, Lemieux I, Akméras N, Bergeron J, et al. (2001) Deterioration of the metabolic risk profile in women. Respective contributions of impaired glucose tolerance and visceral fat accumulation. Diabetes Care 24: Fujimoto WY, Bergstrom RW, Boyko EJ, Chen KW, Leonetti DL, et al. (1999) Visceral adiposity and incident coronary heart disease in Japanese-American men: The 10-year follow-up results of the Seattle Japanese- American Community Diabetes Study. Diabetes Care 22: Boyko EJ, Fujimoto WY, Newell-Morris L (2000) Visceral adiposity and risk of type 2 diabetes: a prospective study among Japanese Americans. Diabetes Care 23: Hayashi T, Boyko EJ, Leonetti DL, McNeely MJ, Newell-Morris L, et al. (2003) Visceral adiposity and the prevalence of hypertension in Japanese Americans. Circulation 108: Simajiri Y, Tsunoda K, Furuta M, Kadoya Y, Yamada S, et al. (2008) Prevalence of metabolic syndrome in Japanese type 2 diabetic patients and its significance for chronic vascular complications. Diabetes Res Clin Pract 79: Yasuda T, Matsuhisa Y, Fujiki N, Sakamoto F, Tsuji M, et al. (2007) Is central obesity a good predictor of carotid Atherosclerosis in Japanese type 2 diabetes with metabolic syndrome? Endocr J 54: Ministry of Health, Labour and Welfare Home page. 21 Nov < html> [Accessed 28 Feb 2017] (In Japanese). 11. Okauchi Y, Iwahashi H, Okita K, Funahashi T, Kishida K, et al. (2013) Weight reduction is associated with improvement of glycemic control in Japanese men, whose hemoglobin A1C is %, with visceral fat accumulation, but not without visceral fat accumulation. J Diabetes Investig 4: Iwahashi H, Noguchi M, Okauchi Y, Morita S, Imagawa A, et al. (2015) Extent of weight reduction necessary for minimization of diabetes risk in Japanese men with visceral fat accumulation and glycated hemoglobin of %. J Diabetes Investig 6: The Examination Committee of Criteria for Metabolic Syndrome (2005) Definition and criteria of metabolic syndrome. J JPN Soc Int Med 94: (In Japanese). 14. Marzetti E, Calvani R, Tosato M, Cesari M, Di Bari M, et al. (2017) Sarcopenia: an overview. Aging Clin Exp Res 29: Examination Committee of Criteria for Obesity Disease in Japan; Japan Society for the Study of Obesity (2002) New criteria for obesity disease in Japan. Circ J 66: Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, et al. (2009) Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 120: Matsuzawa Y (2005) Metabolic syndrome-definition and diagnostic criteria in Japan. J Artheroscler Thromb 12: The Committee of the Japan Diabetes Society on the Diagnostic Criteria of Diabetes Mellitus, Seino Y, Nanjo K, Tajima N, Kadowaki T, et al. (2010) Report of the committee on the classification and diagnostic criteria of diabetes mellitus. J Diabetes Investig 1: Kashiwagi A, Kasuga M, Araki E, Oka Y, Hanafusa T, et al. (2012) International clinical harmonization of glycated hemoglobin in Japan: From Japan Diabetes Society to National Glycohemoglobin Standardization Program values. J Diabetes Investig 3: Nishimura A, Sawai T (2006) Determination of adiponectin in serum using a latex particle-enhanced turbidimetric immunoassay with an automated analyzer. Clin Chim Acta 371: Gautier A, Roussel R, Ducluzeau PH, Lange C, Vol S, et al. (2010) Increases in waist circumference and weight as predictors of type 2 diabetes in individuals with impaired fasting glucose: influence of baseline BMI: data from the DESIR study. Diabetes Care 33: Ryo M, Funahashi T, Nakamura T, Kihara S, Kotani K, et al. (2014) Fat accumulation and obesity-related cardiovascular risk factors in middle-aged Japanese men and women. Intern Med 53: Kanai H, Tokunaga K, Fujioka S, Yamashita S, Kameda- Takemura KK, et al. (1996) Decrease in intra-abdominal visceral fat may reduce blood pressure in obese hypertensive women. Hypertension 27: Okauchi Y, Nishizawa H, Funahashi T, Ogawa T, Noguchi M, et al. (2007) Reduction of visceral fat is associated with decrease in the number of metabolic risk factors in Japanese men. Diabetes Care 30: