The prevalence rate of childhood obesity has been increasing throughout the world in developed countries as well as low- and middle-income countries. What is worse is that the relative increasing rate of childhood obesity is faster than that of adult obesity [1,2]. As childhood obesity has increased, many children have been exposed to the risk of metabolic complications of obesity including type 2 diabetes [3]. In addition, type 2 diabetes in Asia has been rapidly increasing in recent decades and is characterized by onset at younger age and lower body mass index (BMI) compared to in the West [4]. Insulin resistance (IR) is an early indicator of impaired glucose metabolism in type 2 diabetes [5] and pre-teen IR can predict future impaired fasting glucose and type 2 diabetes [6]. Therefore, IR is an excellent clinical marker for identifying high risk children for diabetes prevention interventions. Homeostasis model assessment of insulin resistance (HOMA-IR) has been reported to be reliable and valid as a simple and practical surrogate marker to assess IR among children and adolescents [7,8,9].

Several factors including family history of type 2 diabetes and IR have been identified as risk factors of IR among children and adolescents [10]. Pankow et al. [11] reported that IR was greater in children of parents with type 2 diabetes or insulin resistance syndrome (IRS). Sex [12], race [13], pubertal stage [14] and degree of adiposity [15] have been studied as demographic and physiological factors related to IR in children. It has been also reported that lifestyle factors like physical activity [16] and fruit intake [17] and socio-economic factors like parental education level [18] are associated with IR in children. However, these results have been derived from studies on the Western children. To our knowledge, there have been few studies about related factors of IR in Asian children.

This study was conducted to investigate parental, physiological, behavioral and socio-economic factors related to IR in Korean children.

Children’s mean age was 7.3 ± 0.3 years. Some anthropometric variables such as BMI and waist circumference were higher in boys than girls. However, there were no statistical differences of glucose metabolism indices such as fasting glucose, fasting insulin and HOMA-IR between boys and girls. Children’s mean levels of fasting glucose, fasting insulin and HOMA-IR were 89.1 ± 7.1 mg/dL, 5.04 ± 1.83 μU/mL and 1.11 ± 1.86, respectively (Table 1). Boys reported more frequent moderate and vigorous physical activities compared to girls. The other lifestyle habits were not different between boys and girls (data not shown).

The mean ages of mothers and fathers were 37.7 ± 3.9 years and 40.5 ± 4.2 years, respectively. Mother’s mean levels of fasting glucose, fasting insulin and HOMA-IR were 87.6 ± 7.3 mg/dL, 3.40 ± 2.22 μU/mL and 0.73 ± 2.27, respectively. Father’s mean levels of fasting glucose, fasting insulin and HOMA-IR were 94.1 ± 15.7 mg/dL, 10.17 ± 2.19 μU/mL and 2.34 ± 2.32, respectively (Table 2). It was investigated that more than 70.0% of parents graduated from college and 34.9% of mothers were employed. In addition, 36.9% of families had a monthly household income over 4 million won (exchange rate in 2005: 1,013 Korean Won for 1 US dollar; data not shown).

Sex-adjusted partial correlation coefficients between children’s glucose metabolism indices, and children’s adiposity indices and glucose metabolism indices of their parents were presented in Table 3. All children’s adiposity indices such as waist circumference, BMI and BF% were positively correlated with children’s levels of fasting insulin (r = 0.325–0.338, all P-value < 0.01) and HOMA-IR (r = 0.309–0.318, all P-value < 0.01). Whereas maternal levels of fasting insulin and HOMA-IR have significant positive correlation with children’s level, no significant correlation of levels of fasting insulin and HOMA-IR between fathers and children was not observed. Children’s fasting glucose level had no statistical significant correlations with any of children’s adiposity indices. However, children’s fasting glucose level were positively correlated with parental fasting glucose levels (mothers: r = 0.270, fathers: r = 0.264; both P-value < 0.01). Significant correlations between parental levels of adiposity indices and children’s levels of fasting insulin and HOMA-IR were not observed (data not shown).

Sex-adjusted mean levels of fasting glucose, fasting insulin and HOMA-IR according to children’s lifestyle habits were shown in Table 4. Children’s lifestyle habits related with fasting insulin or HOMA-IR was not identified. In addition, there were no statistical differences of mean levels of fasting glucose, fasting insulin and HOMA-IR according to SES of families and children’s birth weight (data not shown).

Multiple regression analyses were conducted to identify factors that were independently related to children’s HOMA-IR. Sex, waist circumference and maternal level of HOMA-IR were included as independent variables in the main regression model. Children’s waist circumference and maternal level of HOMA-IR had significant positive relationships with children’s HOMA-IR level. An increase of 1 percentage point in body fat was related to 2.7% increase in children’s HOMA-IR (P-value < 0.001) and an increase of 1% of maternal HOMA-IR level was related to 0.2% increase in children’s HOMA-IR (P-value = 0.002) (Table 5). In case waist circumference or BMI was in place of BF% in a regression model, there was no change of the results (data not shown). These results were also consistent in analysis where fasting insulin level was substituted as dependent variable for level of HOMA-IR (data not shown).

This study aimed to investigate which factors were associated with IR in Korean children using cross-sectional data from 111 children aged 7 years and their parents. We thought that parental, physiological, behavioral and socio-economic factors have their own role for the development of IR in children. This study shows that maternal IR and amount of adiposity have significant independent relationships with children’s IR.

There are a lot of evidences that impaired glucose metabolism begins early in life [21,22]. However, fasting glucose in children may not be a proper clinical marker because impaired fasting glucose is a late indicator of impaired glucose metabolism [23] and fasting glucose in children has remarkably narrow distribution [24]. Significant difference of fasting glucose level according to amount of adiposity or lifestyle habits is not observed in this study, as well. On the contrary, IR in children is an early indicator of impaired glucose metabolism [5] and a predictor of future impaired fasting glucose and type 2 diabetes [6]. Various methods were developed for assessing IR including the hyper-insulinemic-euglycemic clamp [25] and the frequently sampled intravenous glucose tolerance test [26]. However, these methods are too invasive, complicated and expensive for epidemiologic study and routine clinical assessment of children. HOMA-IR was developed as a simple and practical surrogate marker to assess IR [20] and has been proved to be reliable and valid among children and adolescents in several studies [7,8,9].

This study shows that HOMA-IR level is higher in children whose mothers have higher HOMA-IR level. Children’s IR is known to have familial background [27]. Children of parents with type 2 diabetes or IRS have higher IR [11]. There are a lot of study results on heritability of HOMA-IR or fasting insulin. Although there is a little difference of study results according to study population and methods, it has been reported that heritability of HOMA-IR or fasting insulin is from 0.23 to 0.54 [28,29,30,31,32,33]. In this study, only IR of mothers among two parents had the significant relationship with children’s IR and IR of fathers was not related to children’ IR. This difference in the relationship with children’s IR between IRs of both parents could not be explained by genetics. Freeman et al. [28] reported that childhood household environment significantly contribute to variance of HOMA-IR level in adulthood. Parents share not only genes but also environments with their children. As mothers primarily bring up their children in the Korean society, they might share more household environment and health behaviors with their children. The stronger linkage between mothers and children could be an important cause that maternal IR was different from paternal IR in the relationship with children’s IR.

All children’s adiposity indices such as waist circumference, BMI and BF% are positively correlated with children’s HOMA-IR level in this study. This result is consistent with previous studies. Previous studies have reported that obesity is one of the most important determinants of IR in children. Lee et al. [12] investigated the distribution and determinants of HOMA-IR among U.S. adolescents who participated in the National Health and Nutrition Examination Survey 1999–2002. They reported that obese children have significantly higher HOMA-IR level compared to normal weight children (mean HOMA-IR level: 4.93 vs. 2.30). Young-Hyman et al. [34] demonstrated that HOMA-IR is positively correlated with amount of adiposity in 5–10 years old overweight children. Additionally, Gardner et al. [35] showed that adiposity at 5 years can predict HOMA-IR at 8 years, especially in girls.

We could not find children’s lifestyle habits related to children’s HOMA-IR in this study. It is well known that regular physical activity can improve insulin sensitivity [36]. It is recommended that at least 60–90 minutes of daily moderate-vigorous physical activity is required for prevention and management of type 2 diabetes in youth [37]. Even though children’s levels of fasting insulin and HOMA-IR decreased with increase of vigorous physical activity frequency in this study, this was not statistically significant. This insignificant result could be because of inaccuracy of physical activity measurement. We measured physical activity not by objective tools like accelerometers but by self-administered questionnaires. Although it was well studied that unhealthy eating habits are associated with adiposity, an important determinants of IR in children [38], there are few studies about their direct relationship with children’s IR. Lindquist et al. [17] reported that fruit and carbohydrate intakes are positively associated with insulin sensitivity and vegetable intake is negatively associated with acute insulin response in African American and white children. However, we could not identify any relationship between children’ eating habits and HOMA-IR. We think that further research is needed to examine dietary factors associated with IR in children.

We reported a result on the relationship between SES of families and children’s overweight in the Korean society in a previous study [39]. The study showed that low maternal education level was related to overweight of children aged 8 years in Korean society. Because overweight is one of the most important determinants of IR in children, we expected to identify the relationships between children’s IR and SES of families including maternal education level in this study. However, we could not find any relationship between SES of families and children’s HOMA-IR in this study. There were few previous studies on the relationship between SES of families and children’s IR. Contrary to our results, Goodman et al. [18] reported that lower parent education was associated with higher IR in adolescents aged 12–19 years. It is difficult to compare the results of both studies because participants of both studies are different in ages. The relationships between parental education level and children’s IR could be changing depending on children’s ages and the influence of parental education level on children’s IR could appear later than that on children’s adiposity.

This study has several limitations to be considered. Cross-sectional design of this study prevented causal inferences. We measured IR not by more precise clamp studies, but by HOMA-IR. Clamp studies cannot be used because of invasiveness, cost and feasibility. HOMA-IR was proved to correlate with more precise measures of IR [7,8,9]. This study is also limited because children’s lifestyle habits were assessed only by self-administered questionnaires. Even though someone can be interested in which factors had the largest influence on children’s IR, this study could not reveal exactly the extent that each factor affected children’s IR. In spite of these several limitations, this study has advantages in directly measuring adiposity and HOMA-IR in both children and their parents. To our knowledge, this study is the first study that investigates parental, physiological, behavioral and socio-economic factors associated with IR in Asian children.

Increased adiposity makes type 2 diabetes develop in children with familial predisposition in younger age. We demonstrate that adiposity and maternal IR are independently associated with children’s IR. Prevention and management of obesity could be practical targets for prevention of type 2 diabetes in Korean children.