2023 Volume 70 Issue 4 Pages 445-452
The aim of the study was to determine the pathogenesis of non-obese children with type 2 diabetes, and its relationship with fat distribution. The study participants included 36 obese children with type 2 diabetes (age: 13.5 years, BMI: 28.3, BMI percentile: 91.9) and 30 non-obese children with type 2 diabetes (age: 13.5 years, BMI: 23.1, BMI percentile: 74.0). The proportion of female participants was significantly higher in non-obese children than in obese children (73.3% vs. 41.7%, p < 0.001). Abdominal fat distribution, evaluated by subcutaneous fat (SF) area, visceral fat (VF) area, and the ratio of VF area to SF area (V/S ratio), measured using computed tomography, and serum lipid levels and liver function were compared between the two groups. Non-obese children with type 2 diabetes had significantly smaller SF area and also smaller VF area than obese children with type 2 diabetes (SF area: 158.3 m2 vs. 295.3 m2, p < 0.001, VF area: 71.0 m2 vs. 94.7 m2, p = 0.032). Whereas non-obese children with type 2 diabetes had significantly greater V/S ratio than obese children with type 2 diabetes (0.41 vs. 0.31, p = 0.007).The prevalence of dyslipidemia and liver dysfunction were similar in the two groups. In conclusion, non-obese children with type 2 diabetes had excess accumulation of VF despite a small amount of SF, which might be associated with glucose intolerance and other metabolic disorders.
THE DIAGNOSIS of type 2 diabetes was previously restricted to adults, particularly in populations of European descent. Type 2 diabetes is now widely recognized to occur in childhood, and the number of children diagnosed with type 2 diabetes has greatly increased worldwide [1, 2]. The most common feature of type 2 diabetes in children is obesity, as in adults. In the United States, almost all have body mass index (BMI) above the 85th percentile for age and sex [3], with the median BMI above the 99th percentile. In Europe, nearly half of adolescents with type 2 diabetes are extremely obese with a BMI above the 99.5th percentile [4]. Abdominal fat distribution is an important factor in the pathogenesis of type 2 diabetes. Greater accumulation of abdominal visceral fat (VF) has been reported to be associated with decreased insulin sensitivity [5-7]. Excess accumulation of VF decreases sensitivity to insulin stimulation of glucose uptake [8], reduces free fatty acid (FFA) reesterification [9], and increases resistance of lipolysis to the inhibitory effect of insulin [10]. Gastaldelli et al. [11] demonstrated that VF, and not subcutaneous fat (SF), accumulation was associated with decreased insulin sensitivity and enhanced gluconeogenesis in individuals with type 2 diabetes. In a previous study, we found that the abdominal fat distribution in children with type 2 diabetes differed from that of children with simple obesity, and that VF accumulation might be an important risk factor for the development of hyperglycemia [12]. While fat distribution, represented by the VF area and SF area, is an essential determinant of insulin resistance [13], several studies have shown that, compared with the VF area alone, the ratio of VF area to SF area (V/S ratio) is more strongly related to metabolic disorders [12, 14, 15].
In contrast, some children with type 2 diabetes, particularly children of Asian descent, are not obese. In Japan, 10–15% of children with type 2 diabetes are not obese [16, 17]; and half of South Asian urban children [18], Taiwanese children [19], and Indian urban children [18] with type 2 diabetes are not obese. Yoon et al. [20] reported that despite the average BMI in Asian children being lower than that of European children, Asian children had higher prevalence of type 2 diabetes. As a possible explanation, it has been speculated that for the same BMI, Asian children might have a greater amount of VF than children of European descent [21, 22].
We previously reported the clinical characteristics of Japanese non-obese children with type 2 diabetes [17]; however, we did not examine the fat distribution in these patients, which could be associated with the pathogenesis of non-obese type 2 diabetes. In the present study, we studied the abdominal fat distribution, including the VF area, SF area, and V/S ratio, measured using abdominal computed tomography (CT) in non-obese children with type 2 diabetes, and compared it with that of obese children with type 2 diabetes. The aim of the study was to compare the VF distribution of non-obese children with type 2 diabetes with that of obese children with type 2 diabetes, and to determine the impact of fat distribution on the pathogenesis of non-obese type 2 diabetes.
The study population included Japanese children aged 10–15 years, living in the Tokyo Metropolitan Area in Japan. The study recruited children with type 2 diabetes who visited the outpatient clinics of Nihon University Hospital during the period from 2000 to 2020, with clinical symptoms of diabetes, a request for a medical evaluation of obesity from the local clinics or parents, or a request for a detailed medical evaluation because of a positive urine glucose result in the school-based screening program [23, 24]. Study subjects were enrolled after providing informed consent to their parents for undergoing an abdominal CT scan to evaluate their VF distribution.
Study definitionObesity was defined as a BMI exceeding the 90th percentile for sex- and age-matched Japanese children [25]. Diagnosis of diabetes was defined according to the American Diabetes Association criteria [26], i.e., 1) a fasting plasma glucose (FPG) ≥126 mg/dL, a random plasma glucose ≥200 mg/dL, or a 2-hour plasma glucose on an oral glucose tolerance test (OGTT, 1.75 g/kg and max 75 g of glucose) ≥200 mg/dL; 2) hemoglobin A1c (HbA1c) ≥6.5%; and 3) symptoms of diabetes such as polyuria, polydipsia, nocturia, and unexplained weight loss. Type 2 diabetes was diagnosed based on maintaining sufficient insulin secretion and not requiring insulin treatment within at least 2 years after the diagnosis of diabetes, managed using dietary management and physical exercise or oral hypoglycemic drugs. None of the patients with type 2 diabetes tested positive for pancreatic β-cell-associated autoantibodies, including insulin autoantibody, glutamic acid decarboxylase (GAD) antibody, and insulinoma-associated protein-2 (IA-2) antibody at the time of diagnosis. In addition, none of the patients had diabetes-associated genetic disorders, such as maturity-onset diabetes (MODY) in the young or mitochondrial diabetes by genetical analyses [23, 24]. However, in non-obese children it was sometimes initially difficult to distinguish between type 2 diabetes and slowly progressive type 1 diabetes. In these cases, it was necessary to follow up the residual endogenous insulin secretion and β-cell autoimmunity during the course of diabetes. Patients with slowly progressive type 1 diabetes usually lost β-cell function within 2 to 3 years of the diagnosis and showed positive results for β-cell associated autoantibodies during the course of diabetes [27, 28]. Whereas patients with type 2 diabetes showed negative results for any pancreatic β-cell-associated autoantibodies during the course of diabetes and maintain sufficient β-cell function at least over 3 years after diagnosis without requiring intensive insulin treatment [17].
Study designThe SF area, VF area, and V/S ratio values measured using abdominal CT, conducted at the time of diagnosis or within 2 months after the diagnosis, were compared between children with non-obese and obese type 2 diabetes. In addition, the FPG, HbA1c, and serum total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride levels at the time of diagnosis with a fasting state were compared between the two diabetic groups. No participants received pharmacological therapies including oral hypoglycemic drugs and insulin at the time of measurement of the SF area, VF area, and the V/S ratio, and when the laboratory tests were performed.
MeasurementsHeight, weight, and waist circumference were measured by experienced nurses in each visit at the outpatient-clinic. BMI was calculated as body weight (kg)/body height (m)/body height (m). The BMI percentile was calculated based on the standard weights of sex-, age-, and height-matched participants [25].
The SF area, VF area, and the V/S ratio were measured at the time of diagnosis or within 2 months after the diagnosis. The SF area and VF area were determined by CT at the umbilical level with a single 8-mm slice, which is recognized as the gold standard for measuring intra-abdominal fat volumes [29, 30]. All participants underwent CT scanning in the supine position using an Aquilion CT scanner (Canon Medical Systems Corporation, Tochigi, Japan). The CT scans were performed at the end of the expiratory phase using settings of 120 κV and 130 to 160 mA. The VF area and SF area were calculated using Synapse Vincent, a software program designed by Fujifilm, Tokyo, Japan [31]. The Synapse Vincent algorithm segmented body regions, and found the external and internal regions of the abdominal wall. The external regions were calculated as fat signal areas facing the body surfaces, where fat was considered to display as approximately –200 to –50 Hounsfield units, and the internal regions were calculated as inside the rib cages and vertebrae. The VF and SF volumes were calculated based on the thresholds of the fat signals of the two regions. The V/S ratio was calculated as the VF area divided by the SF area [12].
The homeostasis model assessment-insulin resistance (HOMA-IR) was calculated as FPG (mg/dL) × fasting immunoreactive insulin (IRI) (μU/mL)/405 [32]. Plasma glucose was measured using the glucose oxidase method and HbA1c was measured using high-performance liquid chromatography during the entire study period at Nihon University Hospital. HbA1c levels expressed as the Japan Diabetes Society (JDS) value were converted to the National Glycohemoglobin Standardization Program (NGSP) value (reference range for NGSP value: 4.6–6.2%). The IRI, total cholesterol, triglyceride, aspartate aminotransferase (AST), and alanine aminotransferase (ALT) levels were measured using standard enzymatic methods. HDL cholesterol was measured by a direct method using a combination of chemical-modified enzymes and sulfated a-cyclodextrin.
Statistical analysisThe results were expressed as medians (ranges). The Mann-Whitney U test and Fisher’s exact test were used to assess the significance of differences between two groups. A p-value <0.05 was considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA).
Study approvalThis study was approved by the ethics committee of Nihon University Hospital (No. 20210702, 21 July 2021) and was performed in accordance with the ethical standards set forth in the 1964 Declaration of Helsinki and its later amendments.
During the period from 2000 to 2020, 168 children were diagnosed with having type 2 diabetes, of whom 134 were obese and 34 were non-obese. Among them, 66 with type 2 diabetes, including 36 obese (male/female = 21/15) and 30 non-obese children (male/female = 8/22), were enrolled in the study to perform an abdominal CT scan. Overall, 102 children with type 2 diabetes, including 134 obese (male/female = 65/69) and 4 non-obese children (male/female = 1/3) were excluded from the study, because we could not obtain the informed consent to perform an abdominal CT scan from them.
Clinical characteristics of the participants at the time of diagnosisThe clinical characteristics of the participants at the time of diagnosis are shown in Table 1. The age at diagnosis was 13.5 (10.0–15.0) years in non-obese children, and 13.5 (10.0–15.0) years in obese children. The distribution of Tanner pubertal stages, comprehensively evaluated by findings of breast, pubic hair, and genital, was not different between the two groups. Whereas the proportion of female participants was significantly higher in non-obese children than in obese children (73.3% vs. 41.7%, p < 0.001). The BMI, BMI percentile and waist circumference were 23.1 (17.6–24.6), 74.0 (57.8–78.2), 78.9 (69.0–81.8) cm in non-obese children, and 28.3 (25.4–41.9), 91.9 (87.0–99.9), 95.8 (86.0–112.2) cm in obese children, respectively. No participants were treated with pharmacological therapies at the time of measurement of the SF area, VF area, and the V/S ratio.
| Non-obese children with type 2 diabetes (n = 30) | Obese children with type 2 diabetes (n = 36) | |
|---|---|---|
| Male/Female (proportion of female) | 8/22* (73.3%) | 21/15 (41.7%) |
| Age (years) | 13.5 (10.0–15.0) | 13.5 (10.0–15.0) |
| Tanner stage (number) | stage 2 (1), stage 3 (11), stage 4 (18) | stage 2 (2), stage 3 (14), stage 4 (20) |
| BMI | 23.1 (17.6–24.6) * | 28.3 (25.4–41.9) |
| BMI percentile | 74.0 (57.8–78.2) | 91.9 (87.0–99.9) |
| Waist circumference (cm) | 78.9 (69.0–81.8)* | 95.8 (86.0–112.2) |
BMI: body mass index
*p < 0.001
Non-obese children with type 2 diabetes had significantly smaller SF area and also smaller VF area than obese children with type 2 diabetes (SF area: 158.3 [45.0–327.9] m2 vs. 295.3 [193.0–605.0] m2, p < 0.001, 95% CI: –183.1, –107.3, VF area: 71.0 [21.1–192.4] m2 vs. 94.7 [49.8–206.0] m2, p = 0.032, 95% CI: –40.8, –4.8) (Fig. 1-A, 1-B). Whereas non-obese children with type 2 diabetes had significantly greater V/S ratio than obese children with type 2 diabetes (0.41 [022–1.48] vs. 0.31 [0.18–0.49], p = 0.007, 95%CI: –0.033, –0.17, Fig. 1-C).
Comparison of subcutaneous fat area between non-obese children with type 2 diabetes and obese children with type 2 diabetes
SF area: subcutaneous fat area
Comparison of visceral fat area between non-obese children with type 2 diabetes and obese children with type 2 diabetes
VF area: visceral fat area
Comparison of the visceral-to-subcutaneous fat ratio between non-obese children with type 2 diabetes and obese children with type 2 diabetes
V/S ratio: visceral-to-subcutaneous fat ratio
The IRI level and HOMA-IR were significantly higher in obese children with type 2 diabetes than that in non-obese children with type 2 diabetes (both p < 0.001), but the FPG, HbA1c, total cholesterol, HDL cholesterol, triglyceride, AST, and ALT levels did not differ significantly between the two groups (Table 2).
| Non-obese type 2 diabetes | Obese type 2 diabetes | p value (95% CI) | |
|---|---|---|---|
| FPG (mg/dL) | 124 (103–179) | 125 (94–232) | 0.877 (–10, 14) |
| HbA1c (%) | 8.4 (6.7–12.2) | 8.5 (6.9–15.0) | 0.779 (–0.5, 0.4) |
| IRI | 12.1 (3.8–31.5) | 27.0 (10.6–102.2) | <0.001 (–19.1, –7.7) |
| Total cholesterol (mg/dL) | 180 (129–275) | 164 (126–287) | 0.477 (–28, 14) |
| HDL-cholesterol (mg/dL) | 46 (30–66) | 47 (30–69) | 0.441 (–9, 4) |
| Triglyceride (mg/dL) | 147 (84–190) | 145 (92–265) | 0.500 (–22, 45) |
| AST (IU/mL) | 29 (9–42) | 29 (14–124) | 0.402 (–28, 11) |
| ALT (IU/mL) | 29 (13–74) | 45 (15–239) | 0.148 (–47, 8) |
| HOMA-IR | 3.8 (3.0–14.5) | 9.2 (3.3–44.1) | <0.001 (–6.4, –2.2) |
FPG: fasting plasma glucose, HbA1c: hemoglobin A1c, IRI: immunoreactive insulin, HDL-cholesterol: high-density lipoprotein, AST: aspartate aminotransferase, ALT: alanine aminotransferase, HOMA-IR: homeostasis model assessment-insulin resistance
In this study, we found that non-obese children with type 2 diabetes had a significantly higher V/S ratio than obese children with type 2 diabetes, despite having a significantly lower BMI, waist circumference and smaller SF area and VF area than obese children with type 2 diabetes. We have previously reported that obese children with type 2 diabetes had greater VF area and higher V/S ratio than children with simple obesity, even within the same BMI percentile [12], and speculated that obese children with type 2 diabetes might have increased inflammation in the VF tissue, which could contribute to insulin resistance and lead to hyperglycemia. VF is highly metabolically active and constantly releases FFA and a variety of adipocytokines into the portal vein circulation [33, 34]. Consequently, VF plays a major role in the development of metabolic syndrome, including promoting hyperinsulinemia, dyslipidemia, and atherosclerosis [35]. VF is reported to produce several proinflammatory factors; while the production of anti-inflammatory cytokines and adipokines are decreased, which contributes to a subclinical systemic inflammation, which may be directly associated with the metabolic disorders [36].
In this study, a novel finding is that the V/S ratio was higher, despite SF area and VF area being smaller, in non-obese children with type 2 diabetes. Fig. 2 shows the comparison of typical abdominal fat distribution in non-obese children with and obese children with type 2 diabetes. Several studies have shown that the V/S ratio is more strongly associated with metabolic disorders, including type 2 diabetes, than VF alone or BMI [12, 14, 15, 33, 37-40]. Chen et al. [40] reported that SF was associated with a lower risk of developing diabetes in women, and that SF tissue might have protective properties against insulin sensitivity. The exact mechanism of the favorable effect of SF on diabetes risk remains unclear, but several possible mechanisms have been proposed, including the higher secretion of favorable adipocytokines, such as leptin and adiponectin, in SF tissue than in VF tissue [41]. In addition, when SF tissue failed to store excess energy, energy could alternatively be deposited in VF tissue [33]. These finding suggest that excess accumulation of VF, but not SF, may contribute to decreased insulin sensitivity and the development to glucose intolerance. Therefore, the V/S ratio might be a more useful marker than VF alone for evaluating the risk of developing type 2 diabetes. The present study has suggested that non-obese children with type 2 diabetes are likely to have a high V/S ratio, with excess accumulation of VF with less storage of SF, and might be metabolically in a more inflammatory state and produce more adipocytokines, leading to glucose intolerance. We have previously demonstrated that non-obese children with type 2 diabetes showed impaired endogenous insulin secretion compared with obese children with type 2 diabetes. They tended to have lower residual β-cell function and required pharmacological therapy, particularly insulin treatment, at an earlier stage of diabetes [17]. Therefore, the combination of insulin secretion failure and insulin resistance associated with excess VF, despite less storage of SF, might contribute to the pathogenesis in non-obese children with type 2 diabetes.
Comparison of typical abdominal fat distribution in obese children with type 2 diabetes and non-obese children with type 2 diabetes
A. A case of non-obese type 2 diabetes, 15 years old, female. BMI 22.4, waist circumference 80.6 cm, SF area 95.8 m2, VF area 117.0 m2, V/S ratio 1.22.
B. A case of obese type 2 diabetes, 14 years old, female. BMI 28.5, waist circumference 87.0 cm, SF area 237.0 m2, VF area 66.0 m2, VF/SF ratio 0.28.
It has been reported that in Caucasian populations most children with type 2 diabetes are obese, and also the incidence is higher in males than in females [2]. Contrary to the Caucasian study, we previously demonstrated that non-obese children were not rare in Japanese type 2 diabetes and female was dominant in non-obese type 2 diabetes [17] . Some studies have shown that females are likely to have a higher V/S ratio than males [40, 41]. Meanwhile, Yoon et al. [20] found that Asians had a higher prevalence of type 2 diabetes than Europeans, despite a lower of prevalence of obesity, suggesting Asians are likely to develop glucose intolerance though they are non-obese. Sex and some ethnic factors [21, 22] might contribute to the development of type 2 diabetes in non-obese individuals.
A variety of studies have shown an association between VF accumulation and hyperlipidemia and liver dysfunction [5-7, 14, 15, 33, 37-40]. We have previously reported that children with obese type 2 diabetes had greater VF storge and exhibited higher levels of serum total cholesterol and ALT, and lower levels of HDL cholesterol than children with simple obesity despite similar BMI [12]. A novel finding of this study is that non-obese children with type 2 diabetes had dyslipidemia and liver dysfunction on the same level with obese children with type 2 diabetes despite less storage of SF. Excess accumulation of VF, but not SF, which produces metabolically a variety of harmful adipocytokines, may contribute to these metabolic disorders as well as dysglycemia in non-obese children with type 2 diabetes. Therefore, we emphasize that high V/S ratio was a significant marker for developing metabolic disorders and diabetes in non-obese children with type 2 diabetes.
This study has some limitations. First, it was a single-center, retrospective observational study with a small sample size. In addition, only Japanese children were studied. Therefore, it is necessary to validate the results by conducting a multicenter study with a larger number of participants in several Asian countries with a high incidence of type 2 diabetes despite a low incidence of obesity [21]. Second, the implementation rates of abdominal CT examination in obese children with type 2 diabetes was lower than those in non-obese children with type 2 diabetes. The explanation of the purpose of an abdominal CT scan might have been insufficient in some obese children with type 2 diabetes, and they may not have sufficiently understood the necessity of the CT scan examination, which can be the reason for a low rate of the examination. While we enhanced the significance of the CT examination in non-obese children with type 2 diabetes to clarify the cause of dysglycemia despite normal BMI, which may contribute a high rate of the examination in non-obese children. Third, it might be necessary to compare SF area, VF area, and the V/S ratio between non-obese children with type 2 diabetes and non-obese healthy children to show the significance of V/S ratio in non-obese children with type 2 diabetes. However, we did not have the data in non-obese children without diabetes, because it was very difficult to obtain the informed consent for undergoing an abdominal CT scan to evaluate the VF distribution. It is generally considered that healthy non-obese children do not have excess accumulation of VF (VF <60 m2) and rarely have metabolic disorders [42]. On the other hand, non-obese children who will progress to diabetes might have greater accumulation of VF despite of a small amount of SF even at the prediabetic period.
In conclusion, we found that non-obese children with type 2 diabetes had a significantly higher V/S ratio than obese children with type 2 diabetes, despite having a significantly lower BMI, and significantly smaller waist circumference and SF area than obese children with type 2 diabetes. Non-obese children with type 2 diabetes also had dyslipidemia and liver dysfunction on the same level with obese children with type 2 diabetes. From these findings, the excess accumulation of VF might play a major role for glucose intolerance and other metabolic disorders in non-obese children with type 2 diabetes.
T.U. has received honoraria from Novo Nordisk Pharma Ltd., Eli Lilly Japan K.K., Abbott Japan L.L.C., Terumo Corp., and JCR Pharmaceuticals Co., Ltd. I.M. has received honoraria from MSD Co., Ltd., Novo Nordisk Pharma Ltd., and AbbVie L.L.C. Other authors declare no conflicts of interest.
