Association of birth weight with risk of diabetes mellitus in adolescence and early adulthood: analysis of the Indonesian Family Life Survey

Ratu Ayu Dewi Sartika; Fathimah Sulistyowati Sigit; Edy Purwanto; Norliyana Aris; Avliya Quratul Marjan; Wahyu Kurnia Yusrin Putra; Sutanto Priyo Hastono

doi:10.6065/apem.2346146.073

Ann Pediatr Endocrinol Metab > Volume 28(4); 2023 > Article

Sartika, Sigit, Purwanto, Aris, Marjan, Putra, and Hastono: Association of birth weight with risk of diabetes mellitus in adolescence and early adulthood: analysis of the Indonesian Family Life Survey

Original Article

Annals of Pediatric Endocrinology & Metabolism 2023;28(4): 267-274.

Published online: December 31, 2023

DOI: https://doi.org/10.6065/apem.2346146.073

Association of birth weight with risk of diabetes mellitus in adolescence and early adulthood: analysis of the Indonesian Family Life Survey

Ratu Ayu Dewi Sartika¹

, Fathimah Sulistyowati Sigit^1,²

, Edy Purwanto³

, Norliyana Aris⁴

, Avliya Quratul Marjan⁵

, Wahyu Kurnia Yusrin Putra¹

, Sutanto Priyo Hastono⁶

¹Department of Public Health Nutrition, Faculty of Public Health, Universitas Indonesia, Depok, Indonesia

²Metabolic Disorder, Cardiovascular Diseases, and Aging Research Cluster, Indonesia Medical Education and Research Institute – Faculty of Medicine, Universitas Indonesia, Jakarta, Indonesia

³Department of Economics, Faculty of Economics and Business, Airlangga University, Surabaya, Indonesia

⁴Faculty of Food Science and Nutrition, Universiti Malaysia Sabah, Sabah, Malaysia

⁵Faculty of Health Sciences, Universitas Pembangunan Nasional Veteran Jakarta, Jakarta, Indonesia

⁶Department of Biostatistics and Population, Faculty of Public Health, Universitas Indonesia, Depok, Indonesia

Address for correspondence: Fathimah Sulistyowati Sigit Department of Public Health Nutrition, Faculty of Public Health, Universitas Indonesia, Depok, Indonesia Email: fathimah10@ui.ac.id

Received June 9, 2023 Revised August 13, 2023 Accepted August 26, 2023

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Purpose

We aimed to investigate the association of birth weight with the risk of diabetes mellitus in adolescence and early adulthood in the Indonesian population.

Methods

This study analyzed data from the Indonesian Family Life Survey, a longitudinal study of the Indonesian population with repeated measurements at 3 time points (1997, 2007, and 2014). The subjects observed were children aged 0–59 months in 1997, who were 10–15 years old in 2007, and 17–22 years in 2014. We performed a generalized linear model to investigate the association between birth weight at baseline and the level of hemoglobin A1c (HbA1c) at the 2 follow-up periods. We adjusted the association for the characteristics of the children, parents, and household.

Results

The mean±standard deviation level of HbA1c was 7.35%±0.95% in 2007 and decreased to 5.30%±0.85% in 2014. The crude β (95% confidence interval [CI]) of the association between birth weight and HbA1c was 0.150 (-0.076, 0.377) in 2007 and 0.146 (-0.060, 0.351) in 2014. After adjustment for the sociodemographic characteristics of the children, parents, and confounding factors, the adjusted β (95% CI) was 1.12 (0.40–1.85) in 2007 and 0.92 (0.35–1.48) in 2014. The HbA1c of the parents, father’s employment status, percentage of food expenditure, and underweight were the covariates that had significant associations with HbA1c.

Conclusions

HbA1c level was higher in adolescence than in early adulthood. Birth weight was associated with HbA1c level in both periods. The HbA1c of the parents, father’s employment, percentage of food expenditure, and underweight partly explained the association between birth weight and the HbA1c level.

Keywords: Birth weight, Diabetes mellitus, Adolescence, Early adulthood, Hemo globin A1c, Cohort

Highlights

· Birth weight was positively associated with hemoglobin A1c (HbA1c) levels during adolescence and young adulthood.

· HbA1c level was higher during adolescence than in young adulthood.

· Besides genetics, sociodemographic factors play a part in the level of HbA1c.

Introduction

Diabetes mellitus (DM) is a global health problem, particularly in developing countries, where the prevalence is steadily increasing and is estimated to continue rising until 2035. In Indonesia, according to the 2018 National Health Survey report, the prevalence of DM in the population aged ≥15 years was 2.0% [1]. This rate had increased compared to that documented in the 2013 National Survey, which was 1.5%. The 2013 survey reported that 36.6% of the Indonesian population had impaired fasting blood glucose and was at risk of DM [2].

The 'fetal origins hypothesis' is associated with developmental health, disease, and environmental status of the individual during the fetal period and affects their quality of health and incidence of degenerative diseases in adulthood. It is well established that pregnant women with poor nutritional status have small fetuses. If the fetus who is exposed to intrauterine hunger is born with low birth weight, this may lead to compensatory adaptation mechanisms in the outer-uterine environment after birth. This incompatibility may become a risk factor for noncommunicable diseases, including type 2 DM, hypertension, obesity, dysfunction of lipid metabolism, and cardiovascular disease. An incompatibility may result that can become a risk factor for noncommunicable diseases, including type 2 DM, hypertension, obesity, dysfunction of lipid metabolism, and cardiovascular disease [3]. Extensive epidemiological studies have shown a linear relationship between birth weight and obesity and noncommunicable diseases later in life. In a study with various populations and age groups, it was shown that individuals born with low and very high birth weights are at risk of developing DM in adulthood [4].

One of the parameters for measuring DM is hemoglobin A1c (HbA1c), which measures glucose metabolism in the blood, indicating the average blood glucose level over a period of 2 to 3 months. Metabolically, in its presence, glucose with react with hemoglobin and other proteins to form glycolytic compounds. An advantage of measuring HbA1c in diabetic individuals is its higher preanalytical stability compared to other diagnostic tests for diabetes, which are fasting plasma glucose (FPG) and oral glucose tolerance test (OGTT) [5].

In a previous study conducted in Indonesia (the Tanjung Sari cohort study), a history of low birth weight was associated with impaired glucose tolerance in adolescents aged 12–14 years. The study recommended that the prevention of DM should focus on improving the nutritional status of mothers and newborns [6]. Nevertheless, whether impaired glucose tolerance continues until adulthood and is reflected in the HbA1c level remain unknown. Thus, in the present study, we investigated the association between infant birth weight and HbA1c level in adolescence and early adulthood.

Materials and methods

1. Data collection

This study used secondary data from the longitudinal study of the Indonesia Family Life Survey (IFLS). The measurements were obtained at 5 time points: 1993, 1997, 2000, 2007, and 2014. According to the IFLS data user manual, a sample of 7,224 households was obtained from 13 of 26 provinces in Indonesia, which were located on the islands of Java, Sumatra, Bali, West Nusa Tenggara, Kalimantan, and Sulawesi. The population selected as the IFLS data sample represented approximately 83% of the total Indonesian population. The survey recorded data on fertility, family planning and contraception, infant and child health, education, migration, employment, and socioeconomic and health status of the population [7]. The survey provided data at the individual and household levels. The information was linked to the local health facility and community, enabling investigation of the impact of public policies on the lives of respondents, as well as the impact of social, economic, and environmental changes on population health [8].

This observational study used longitudinal data from IFLS. Data collection began when the children were aged 0–59 months in 1997, and they were followed until they were aged 10–15 years in 2007, and 17–22 years in 2014. The households included had complete data on birth weight and HbA1c values at all study time points. In addition, we only included children with complete data on parental and household characteristics.

2. HbA1c

According to the American Diabetes Association and the World Health Organization (WHO), the HbA1c level is a criterion in the diagnosis of DM [9]. In this study, which employed data from IFLS, the HbA1c levels were measured with the dried blood spot (DBS) method, involving collection of a blood drop, usually from the fingertip, onto a special paper and then allowing it to dry before analysis. DBS is used in IFLS because it is easier to obtain a sample compared with venous blood samples. Moreover, DBS can be stored long term in stable clotting conditions. HbA1c measurement was chosen for IFLS because the respondents were not required to fast in advance, unlike fasting blood glucose measurement. An individual is diagnosed with DM if they have an HbA1c level ≥6.5% and with pre-DM in the HbA1c range from 5.7% to 6.4% [5].

3. Nutritional status

Body weight, height, and waist circumference were measured based on standardized anthropometric methods by trained operators. The body mass index (BMI) was calculated as weight divided by the square of height (kg/m²). Overweight/obesity in children is defined as a BMI-for-age z-score (BAZ) > 1 standard deviation (SD) from the average. In addition to BMI-for-age, other measures of nutritional status in children were recorded, such as weight-for-age, height-for-age, and weight-for-height z-scores.

In adults, overall obesity is defined according to the WHO BMI reference >24.99 kg/m² [10]. Abdominal obesity is defined as a waist circumference >90 cm in men and >80 cm in women [11]. Birth weight in infants is categorized as low (<2,500 g) or normal (>2,500 g).

4. Sociodemographic and lifestyle variables

The IFLS collected information on the characteristics of children and their parents. The characteristics included sex, age, education (high school graduate or higher or junior high school graduate or lower), occupation, household location (urban or rural), household size (nuclear or extended family), per capita expenditure, and food expenditure.

Physical activity was recorded as the intensity, duration, and frequency of exercise. A moderate-to-intense exercise for at least 30 minutes, 5 days per week or more, was considered physically active. Parental smoking habits were categorized as smoking or not smoking. All data were obtained through interviews using the IFLS questionnaire, which is described elsewhere.

5. Data analysis

Statistical analyses were conducted using Stata 15 (StataCorp LLC, College Station, TX, USA). The dependent variable in this study was the risk of DM in adolescence and early adulthood, as observed by the levels of HbA1c in 2007 and 2014. The primary independent variable was the birth weight of children aged 0–59 months in 1997.

To describe the characteristics of the population, we performed univariate data analysis and presented the characteristics as the mean (SD) or median (interquartile range) for continuous variables or as percentages (%) for categorical variables. We then performed bivariate analysis to examine the associations of each individual, parent, lifestyle, and sociodemographic characteristic with the levels of HbA1c in 2007 and 2014. We included the significant associations/covariates in the final regression model and adjusted them as confounding factors. In addition, as growing evidence showed that growth stunting in childhood was associated with an increased risk of degenerative and metabolic diseases, such as type 2 DM [12-14], we added the interaction term between birth weight and height-for-age z-score in the multivariate regression models. This approach assessed whether low birth weight in combination with stunting (short for age) increases the risk of diabetes.

We performed a generalized linear model to examine the association of birth weight with the level of HbA1c during the 2 follow-up points (2007 and 2014) separately. We created 2 models: model 1, which measured the association between the birth weight at baseline (1997) and the HbA1c when the children were aged 10–14 years (2007), and model 2, which measured the association between the weight at baseline and the HbA1c level at 17–22 years (2014). A P-value <0.1 indicated a tendency, while a P-value <0.05 with a 95% confidence interval (CI) was considered statistically significant.

6. Data permission and ethical clearance

This study received approval from the RAND Corporation, the institute that conducted the survey and owned the IFLS data. The IFLS survey was reviewed and approved by the Institutional Review Board at the University of Gadjah Mada (ethical clearance number s0064-06-01-CR01).

Results

1. Descriptive characteristics

The mean±SD of infant birth weight at baseline measurement in 1997 was 3.19±0.48 kg. The proportion of infants who were born with low birth weight (<2.5 kg) was 4.2%. There were slightly more boys (52.3%) than girls (47.7%) in the study p opulation (Table 1). More detailed nutritional status measurements of the infants in 1997 are provided in Supplementary Table 1.

At the first follow-up measurement in 2007, the mean±SD level of HbA1c was 7.35%±0.95%. The majority of subjects (92.7%) had an HbA1c level higher than the pre-DM cutoff of >5.7%. According to the BMI stratification, 13.6% of individuals in the population were underweight, 75.6% had normal weight, and 10.8% were overweight/ obese. At the second follow-up in 2014, the mean HbA1c level dropped to 5.30%±0.85%. The proportions of subjects who were underweight, normal weight, and overweight were 17.8%, 68.3%, and 13.9%, respectively (Table 1, Supplementary Table 2).

In regard to dietary intake, 79.8% of the participants in 2007 ate at least 3 full meals per day. This number decreased to 67.5% in 2014. The frequency of intake of common food types is shown in Supplementary Table 3. In regard to physical activity, only 25.0% (in 2007) and 32.5% (in 2014) of the participants were considered physically active.

The parent characteristics, including anthropometric measurements (BMI, waist circumference, hip circumference), HbA1c level, and sociodemographic and lifestyle information are outlined in Table 1 and Supplementary Table 2. In 2007, the mean annual household income was Indonesian Rupiah (IDR) 19,065,875±24,466,441, while it was IDR 30,691,791±31,959,653 in 2014. Of these incomes, the proportion that went toward food expenditure was 56.32%±14.24% in 2007 and 49.36%±15.92% in 2014.

2. Association of birth weight with the level of HbA1c in adolescence and early adulthood

Table 2 presents the results from bivariate analysis examining the associations among the characteristics of children, parents, and households with the levels of HbA1c in 2007 (when the children were aged 10–15 years) and in 2014 (when they were aged 17–22 years). Among all associations, we observed that the weight-for-age z-score at birth and the HbA1c level of the parents were significantly associated with the HbA1c level in early adolescence (2007). In addition, we observed that the weight-for-age z-score at birth, food expenditure, and nonfood expenditure were significantly associated with the level of HbA1c in early adulthood (2014). The crude β (95% CI) of the association between birth weight and HbA1c was 0.150 (-0.076, 0.377) in 2007 and 0.146 (-0.060, 0.351) in 2014.

Table 3 displays the results of the generalized linear model examining the associations between infant birth weight (1997) and the levels of HbA1c in adolescence (2007) and early adulthood (2014). Significant covariates from the prior bivariate analysis were included in this multivariate analysis. The associations were adjusted for the characteristics of children and parents, as well as sociodemographic factors as the confounders. All the covariates with significant associations are shown in Table 3.

We observed that a birth weight of 1 kg was associated with an adjusted β (95% CI) of HbA1c of 1.12 (0.40–1.85) in 2007, whereas it was associated with a 0.92 (0.35–1.48) % increase in HbA1c in 2014. From Table 3, we observe that the father's level of HbA1c and state of underweight were the covariates that were positively associated with the HbA1c level in adolescence (2007). Meanwhile, being underweight, father's unemployment, and mother's HbA1c level were positively associated with HbA1c in early adulthood (2014). We also observed that the percentage of food expenditure was inversely associated with HbA1c in 2014.

Discussion

In this study, we investigated the association of infant birth weight with the risk of developing DM later in adolescence and early adulthood. Using a cohort of the Indonesian population, we observed children aged 0–59 months in 1997 and followed them until they were 10–15 years old in 2007 and 17–22 years old in 2014. We used HbA1c as the biomarker of DM risk.

The level of HbA1c was higher when the individuals were adolescents rather than young adults. Birth weight was positively associated with HbA1c in both follow-up measurements in 2007 and 2014. Several factors partly explained the association between birth weight and HbA1c later in life, namely parents’ levels of HbA1c, father's employment status, percentage of food expenditure, and underweight.

Our findings of higher level of HbA1c in adolescence than in early adulthood is consistent with the findings of several prior studies [15-19]. One study observed that HbA1c tends to rise throughout puberty, as more calories are required for organ maturation. This may result in changes in adipose tissue metabolism, which leads to changes in leptin, adiponectin, and insulin levels [20]. This may lead to more rapid glucose metabolism and control over appetite. Our results also support those of a previous study showing the importance of glycemic control in this age group, and that redefinition of normal glycemia and pre-DM in adolescents may be necessary.

Moreover, our results showed that infant birth weight is positively associated with HbA1c levels in adolescence and early adulthood. Nevertheless, we also observed that underweight was positively associated with the level of HbA1c in the 2 periods. These findings aligned with those of earlier studies, which concluded that infants born with either low or very high birth weight have a greater risk of developing DM [4]. Although high birth weight is a more widely recognized risk factor for childhood obesity and earlier onset of type 2 DM, low birth weight and a history of small for gestational age are associated with higher insulin resistance (mean homeostasis model assessment of insulin resistance) and the components of metabolic syndrome [21-23].

The level of HbA1c of parents was also correlated with the HbA1c level of their offspring in adolescence and early adulthood. Although genetics is a known risk factor, and it is well established that the offspring of DM patients have higher HbA1c levels than those of nondiabetic parents [24]. This phenomenon may be attributed to the similar lifestyle (including dietary intake and physical activity) and shared socioeconomic characteristics within the same household.

In regard to sociodemographic confounding factors, we observed that the employment status of the father and the percentage of food expenditure also partly explained the association between birth weight and HbA1c levels in adolescence and early adulthood. These findings partially aligned with those of earlier studies. A previous study observed that fathers' unemployment and low socioeconomic status were closely linked with higher HbA1c and complications of type 1 DM in their children [15,25-27]. Another study showed that satisfactory employment status of the father is positively associated with better glycemic control in their diabetic offspring [28]. This may be partially explained by their dietary patterns, as families with low socioeconomic status tend to resort to less expensive, less nutritious, and higher carbohydrate diets [29]. Meanwhile, affluent households spend more on fruits, vegetables, and other healthy foods, although no difference was found in their spending on unhealthy foods (e.g., sugar-sweetened beverages) [30].

Another study observed that those whose food expenditure included eating outside the home twice a week or more, particularly fried chicken and hamburgers, were associated with an increased risk of type 2 DM [31]. In another study that focused on households with DM, it was shown that households with DM spend more on unhealthy foods, such as oil and dressing, butter, ready meals, and processed red meat [32].

The strength of this study was that it represented the majority of the Indonesian population, enabling us to generalize the findings to the national population. In addition, although several publications in recent years have used IFLS data to identify the incidence of DM in the Indonesian population, these studies were all cross-sectional analyses that only employed data from a single measurement or time point. The present study was the first to use multi-year observations or repeated measurement data from IFLS (from 1997, 2007, and 2014 time points), which may aid in constructing a causal relationship. Nevertheless, several limitations should be mentioned. First, owing to the observational nature of this study, the possibility existed of residual confounding that may have distorted the true exposure-outcome association investigated. Second, owing to the nature of the data collection performed by visiting the homes of the respondents, the HbA1c levels were measured by the DBS method. This method may result in slight overestimation of the HbA1c level. Nevertheless, a previous study showed that a strong correlation did not exist between HbA1c levels from DBS versus venous plasma. Thus, this difference would have minimal importance [33,34]. This study also included participants from rural areas in Indonesia that often had limited laboratory facilities. Thus, the DBS method was the most feasible and practical test to be conducted. Third, owing to the nature of data collection, the possibility existed of a social desirability bias; i.e., the participants may have provided socially expected answers when the lifestyle (dietary intake, physical activity) was assessed. Nevertheless, lifestyle did not appear to explain the association between birth weight and HbA1c; therefore, this aspect would not have altered our results.

Despite these limitations, as cohorts with repeated measurements in metabolic profiles of the Indonesian population remain rare, we believe that this study contributes to the literature and can become a foundation for more extensive research measuring the effect of birth weight on the risk of DM onset later in life.

In conclusion, we observed that the HbA1c level was higher in adolescence than in early adulthood in this cohort of the Indonesian population. Birth weight was associated with HbA1c levels in both periods, and parents’ HbA1c, father’s unemployment, percentage of food expenditure, and underweight partly explained the association between birth weight and HbA1c level.

Supplementary material

Supplementary Table 1 can be found via https://doi.org/10.6065/apem.2346146.073.

Supplementary Table 1.

Nutritional status of children aged 0–59 months in 1997

apem-2346146-073-Supplementary-Table-1.pdf

Supplementary Table 2.

Lifestyle and sociodemographic characteristics in 2007 and 2014

apem-2346146-073-Supplementary-Table-2.pdf

Supplementary Table 3.

Dietary intake of the participants in 2007 and 2014

apem-2346146-073-Supplementary-Table-3.pdf

Notes

Conflicts of interest

No potential conflict of interest relevant to this article was reported.

Funding

This research was funded by the Directorate of Research and Development, Universitas Indonesia under Hibah PUTI 2022 (Grant No. NKB-1286/UN2.RST/HKP.05.00/2022).

Data availability

The data that support the findings of this study can be provided by the corresponding author upon reasonable request.

Author contribution

Conceptualization: RADS, EP, NA, AQM, WKYP, SPH; Data curation: RADS, EP, NA, AQM, WKYP, SPH; Formal analysis: RADS, FSS, EP, NA, AQM, SPH; Funding acquisition: RADS, EP, NA, AQM, WKYP, SPH; Methodology: RADS, EP, NA, AQM, WKYP, SPH; Project administration: RADS, EP, AQM, WKYP; Visualization: RADS, FSS, EP, NA, AQM, SPH; Writing - original draft: FSS; Writing - review & editing: RADS, FSS

Table 1.

Descriptive characteristics of the population

Characteristic	1997	2007	2014
Children’s characteristics
Male sex (%)	52.3	-	-
Education (duration in yr)	-	-	10.78±2.39
Birth weight (kg)	3.19±0.48	-	-
HAZ score	1.69±1.77	-	-
WAZ score	1.35±1.29	-	-
WHZ score -	0.47(1.45	-	-
BAZ score -	0.32±1.48	-	-
HbA1c (%)	-	7.35±0.95	5.30±0.85
Individuals with HbA1c >5.7% (%)	-	93.7	1.8
Father’s characteristics
Education (duration in yr)	-	7.77±4.1	7.85±4.26
BMI (kg/m²)	-	22.55±3.90	23.19±4.0
Waist circumference (cm)	-	80.30±10.83	84.08±11.53
Hip circumference (cm)	-	89.18±8.30	90.95±8.47
HbA1c (%)	-
Mother’s characteristics
Education (duration in yr)	-	6.93±4.12	6.99±4.23
BMI (kg/m²)	-	23.96±4.25	25.09±4.54
Waist circumference (cm)	-	81.51±11.22	84.93±11.09
Hip circumference (cm)	-	92.58±10.12	94.88±9.09
HbA1c (%)	-	7.86±1.15	5.72±1.26
Household characteristics
Annual income (IDR)	-	19,065,875±24,466,441	30,691,791±31,959,653
Food expenditure (%)	-	56.32±14.24	49.36±15.92
Nonfood expenditure (%)	-	43.68±14.24	50.60±15.92
Per capita expenditure (IDR)	-	426,148±372,520	1,225,738±1,210,163
Household size (no. of persons)	-	5.57±1.99	4.19±2.14

Values are presented as mean±standard deviation unless otherwise indicated.

HAZ, height-for-age z-score; WAZ, weight-for-age z-score; WHZ, weight-for-height z-score; HbA1c, hemoglobin A1c; BMI, body mass index; IDR, Indonesian Rupiah.

Table 2.

The associations of children's, parent's, and household characteristics with the levels of HbA1c in 2007 and 2014 (bivariate analysis)

Characteristic	Model 1 (1997 and 2007)			Model 2 (1997 and 2014)
Characteristic	β	95% CI	P-value	β	95% CI	P-value
Children’s characteristics
Birth weight	0.150	-0.076 to 0.377	0.194	0.146	-0.060 to 0.351	0.164
HAZ score	-0.025	-0.088 to 0.037	0.430	-0.042	-0.099 to 0.015	0.147
WAZ score	-0.100	-0.185 to -0.015	0.021*	-0.081	-0.157 to -0.004	0.039*
WHZ score	-0.068	-0.144 to 0.008	0.081	-0.045	-0.115 to 0.015	0.204
Sex (girls vs. boys)	0.060	-0.158 to 0.278	0.590	-0.110	-0.308 to 0.089	0.279
Nutritional status
Overweight vs. underweight	-0.004	-0.448 to 0.440	0.985	0.060	-0.295 to 0.416	0.740
Normal weight vs. underweight	-0.138	-0.459 to 0.183	0.399	-0.146	-0.410 to 0.118	0.279
Physical active (inactive vs. active)	-0.483	-1.199 to 0.232	0.185	0.048	-0.167 to 0.263	0.663
No. of full meals per day (<3 times/day vs. ≥3 times/day)	-0.044	-0.316 to 0.228	0.751	-0.064	-0.279 to 0.151	0.558
Staple foods consumption (day/wk)				0.261	-0.247 to 0.769	0.314
Protein consumption (day/wk)	-0.018	-0.106 to 0.070	0.693	0.007	-0.075 to 0.088	0.874
Vegetable consumption (day/wk)	0.020	-0.027 to 0.067	0.405	0.010	-0.029 to 0.049	0.619
Fruits consumption (day/wk)	-0.045	-0.146 to 0.056	0.382	-0.017	-0.130 to 0.096	0.768
Sugary foods consumption (consumption (day/wk)				-0.024	-0.062 to 0.014	0.209
Father’s characteristics
Education	-0.006	-0.037 to 0.025	0.694	0.017	-0.011 to 0.045	0.242
Employment status (working vs. unemployed)	0.014	-0.653 to 0.680	0.968
Smoking status (smoker vs. nonsmoker)	-0.130	-0.433 to 0.172	0.398	-0.109	-0.398 to 0.179	0.457
No. of full meals per day (<3 times/day vs. ≥3 times/day)	-0.144	-0.451 to 0.162	0.355	-0.215	-0.523 to 0.094	0.172
Physical activity (inactive vs. active)	0.160	-0.151 to 0.472	0.313	0.190	-0.093 to 0.473	0.189
BMI (kg/m²)	0.003	-0.033 to 0.039	0.851	-0.012	-0.045 to 0.021	0.484
HbA1c level (%)	0.246	0.068 to 0.425	0.007*	0.058	-0.082 to 0.199	0.417
Waist circumference (normal vs. obese)	0.061	-0.399 to 0.521	0.796	0.176	-0.122 to 0.474	0.247
Mother’s characteristics
Education (duration in yr)	-0.013	-0.040 to 0.015	0.365	0.011	-0.080 to 0.103	0.806
Employment status (working vs. unemployed)	-0.200	-0.468 to 0.067	0.142	0.054	-0.198 to 0.305	0.676
Smoking status (smoker vs. nonsmoker)	0.117	-0.499 to 0.732	0.710	-0.159	-0.715 to 0.396	0.575
No. of full meals per day (<3 times/day vs. ≥3 times/day)	-0.081	-0.360 to 0.199	0.572	-0.129	-0.367 to 0.110	0.290
Physical activity (inactive vs. active)	-0.096	-0.346 to 0.155	0.453	0.018	-0.205 to 0.242	0.872
BMI	-0.008	-0.036 to 0.020	0.564	-0.007	-0.031 to 0.018	0.590
HbA1c	0.202	0.081 to 0.323	0.001*	0.095	-0.017 to 0.206	0.097
Waist circumference (normal vs. obese)	0.103	-0.280 to 0.486	0.599	0.025	-0.240 to 0.291	0.851
Household characteristics
Living area (urban vs. rural)	-0.125	-0.344 to 0.093	0.261	0.158	-0.047 to 0.363	0.131
Household size (persons)	0.035	-0.020 to 0.090	0.210	-0.027	-0.074 to 0.019	0.245
Food expenditure (%)	0.002	-0.005 to 0.010	0.543	-0.006	-0.013 to 0.000	0.045*
Nonfood expenditure (%)	-0.002	-0.010 to 0.005	0.544	0.006	0.000 to 0.013	0.044*

HbA1c, hemoglobin A1c; β, coefficient; CI, confidence interval; HAZ, height-for-age z-score; WAZ, weight-for-age z-score; WHZ, weightfor-height z-score; BMI, body mass index.

Table 3.

Multivariate analysis - the associations between birth weight and HbA1c levels in adolescence (10–15 years; 2007) and early adulthood (17–22 years; 2014)

Variable	Model 1 (1997 and 2007)
Variable	β	95% CI	P-value
Intercept	1.626	-1.299 to 4.551	0.276
Birth weight	1.122	0.399 to 1.846	0.002^*
Father's HbA1c level	0.302	0.084 to 0.520	0.007^*
Height-for-age (normal vs. stunted)	3.687	0.585 to 6.790	0.020^*
Weight-for-age (normal vs. underweight)	-0.616	-1.167 to -0.065	0.028^*
Birth weight^† height-for-age	-1.119	-2.048 to -0.190	0.018^*
Father's waist circumference (normal vs. obese)	0.171	-0.413 to 0.756	0.565
	Model 2 (1997 and 2014)
	β	95% CI	P-value
Intercept	4.763	2.235 to 7.291	0.000^*
Birth weight	0.917	0.352 to 1.481	0.001^*
Height-for-age (normal vs. stunted)	2.891	0.608 to 5.175	0.013^*
Father's waist circumference (normal vs. obese)	0.362	0.004 to 0.720	0.047^*
Weight-for-age (normal vs. underweight)	-0.451	-0.866 to -0.037	0.033^*
Father's employment status (working vs. unemployed)	-2.240	-3.662 to -0.819	0.002^*
Percentage of food expenditure	-0.035	-0.059 to -0.010	0.005^*
Mother's HbA1c level	0.052	-0.080 to 0.184	0.442
Sugary foods consumption (day/wk)	-0.046	-0.111 to 0.018	0.161
Birth weight^† height-for-age	-0.900	-1.590 to -0.211	0.011^*
Father's occupation percentage of food expenditure	0.028	0.001 to 0.054	0.039^*

HbA1c, hemoglobin A1c; β, coefficient; CI, confidence interval.

^* P<0.05.

^† The interaction term between the 2 variables (birth weight & height-for-age).

References

1. National Institute for Health Research and Development (NIHRD), Ministry of Health, Republic of Indonesia. Laporan Nasional RISKESDAS 2018 [Internet]. Jakarta (Indonesia), NIHRD. 2018;[cited 2023 May 23]. Available from: http://labdata.litbang.kemkes.go.id/images/download/laporan/RKD/2018/L aporan_Nasional_RKD2018_FINAL.pdf.

2. National Institute for Health Research and Development (NIHRD), Ministry of Health, Republic of Indonesia. Laporan Nasional RISKESDAS 2013 [Internet]. Jakarta (Indonesia), NIHRD. 2013;[cited 2023 May 23]. Available from: http://labdata.litbang.kemkes.go.id/images/download/laporan/RKD/2013/Laporan_riskesdas_2013_final.pdf.

3. Yokoyama M, Saito I, Ueno M, Kato H, Yoshida A, Kawamura R, et al. Low birthweight is associated with type 2 diabetes mellitus in Japanese adults: The Toon Health Study. J Diabetes Investig 2020;11:1643–50.

4. Tian JY, Cheng Q, Song XM, Li G, Jiang GX, Gu YY, et al. Birth weight and risk of type 2 diabetes, abdominal obesity and hypertension among Chinese adults. Eur J Endocrinol 2006;155:601–7.

5. American Diabetes Association. 2. Classification and diagnosis of diabetes. Diabetes Care 2014;38(Supplement_1):S8–16.

6. Hikmat P, Bandiara R, Usman S, Susandi E, Diana A, Arifin A, et al. Impaired glucose tolerance among adolescents with low birth weight history: The Tanjungsari Cohort Study in Indonesia. Asia Pac J Clin Nutr 2020;29:120–6.

7. Frankenberg E, Thomas D. The Indonesia Family Life Survey (IFLS): study design and results from waves 1 and 2. Santa Monica (CA): RAND. 2000.

8. John Strauss, Kathleen Beegle, Bondan Sikoki, Agus Dwiyanto, Yulia Herawati, Firman Witoelar. User’s guide for the Indonesia Family Life Survey, wave 3. Santa Monica (CA): RAND. 2004.

9. World Health Organization. Use of glycated haemoglobin (HbA1c) in the diagnosis of diabetes mellitus. Abbreviated Report of a WHO Consultation [Internet]. Geneva (Switzerland), World Health Organization. 2011;[cited 2023 May 4]. Available from: https://apps.who.int/iris/bitstream/handle/10665/70523/WHO_NMH_CHP_CPM_11.1_eng.pdf.

10. World Health Organization, Western Pacific Region. The Asia-Pacific perspective: redefining obesity and its treatment. Geneva (Switzerland): World Health Organization. 2000.

11. Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, et al. 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 2009;120:1640–5.

12. De Lucia Rolfe E, de França GVA, Vianna CA, Gigante DP, Miranda JJ, Yudkin JS, et al. Associations of stunting in early childhood with cardiometabolic risk factors in adulthood. PLoS One 2018;13:e0192196.

13. Grillo LP, Gigante DP. Evidence for the association between early childhood stunting and metabolic syndrome. Preedy V, Patel VB. Handbook of famine, starvation, and nutrient deprivation: from biology to policy [Internet] Cham, Springer International Publishing. 2017;[cited 2023 May 4]. Available from: https://doi.org/10.1007/978-3-319-40007-5_31-1.

14. Koncz V, Geldsetzer P, Manne-Goehler J, Wendt AS, Teufel F, Subramanian SV, et al. Shorter height is associated with diabetes in women but not in men: nationally representative evidence from Namibia. Obesity (Silver Spring) 2019;27:505–12.

15. Miller KM, Beck RW, Foster NC, Maahs DM. HbA1c levels in type 1 diabetes from early childhood to older adults: a deeper dive into the influence of technology and socioeconomic status on HbA1c in the T1D exchange clinic registry findings. Diabetes Technol Ther 2020;22:645–50.

16. Clements MA, Foster NC, Maahs DM, Schatz DA, Olson BA, Tsalikian E, et al. Hemoglobin A1c (HbA1c) changes over time among adolescent and young adult participants in the T1D exchange clinic registry. Pediatr Diabetes 2016;17:327–36.

17. Jon Edqvist, Araz Rawshani, Aidin Rawshani, Martin Adiels, Stefan Franzén, Lena Bjorck, et al. Trajectories in HbA1c and other risk factors among adults with type 1 diabetes by age at onset. BMJ Open Diab Res Care 2021;9:e002187.

18. Hagger V, Hendrieckx C, Cameron F, Pouwer F, Skinner TC, Speight J. Diabetes distress is more strongly associated with HbA1c than depressive symptoms in adolescents with type 1 diabetes: Results from Diabetes MILES Youth-Australia. Pediatr Diabetes 2018;19:840–7.

19. Seo JY, Hwang S sik, Kim JH, Lee YA, Lee SY, Shin CH, et al. Distribution of glycated haemoglobin and its determinants in Korean youth and young adults: a nationwide population-based study. Sci Rep 2018;8:1962.

20. Kelsey MM, Severn C, Hilkin AM, Pyle L, Nadeau KJ, Zeitler PS. Puberty is associated with a rising hemoglobin A1c, even in youth with normal weight. J Pediatr 2021;230:244–7.

21. Martín-Calvo N, Goni L, Tur JA, Martínez JA. Low birth weight and small for gestational age are associated with complications of childhood and adolescence obesity: systematic review and meta-analysis. Obes Rev 2022;23 Suppl 1:e13380.

22. Prinz N, Putri RR, Reinehr T, Danielsson P, Weghuber D, Norman M, et al. The association between perinatal factors and cardiometabolic risk factors in children and adolescents with overweight or obesity: a retrospective two-cohort study. PLoS Med 2023;20:e1004165.

23. Sun D, Wang T, Heianza Y, Huang T, Shang X, Lv J, et al. Birthweight and cardiometabolic risk patterns in multiracial children. Int J Obes (Lond) 2018;42:20–7.

24. Taiwo EO, Thanni LO, Taiwo OP. Patterns of HbA1c levels in normoglycemic offspring of T2 diabetes mellitus patients. Hosp Pract Res 2022;7:46–9.

25. Archinkova M, Konstantinova M, Savova R, Iotova V, Petrova C, Kaleva N, et al. Glycemic control in type 1 diabetes mellitus among Bulgarian children and adolescents: the results from the first and the second national examination of HbA1c. Biotechnol Biotechnol Equip 2017;31:1198–203.

26. Bijlsma-Rutte A, Rutters F, Elders PJM, Bot SDM, Nijpels G. Socio-economic status and HbA(1c) in type 2 diabetes: a systematic review and meta-analysis. Diabetes Metab Res Rev 2018;34:e3008.

27. Kiess W, Galler A, Schmidt A, Nietzschmann U, Neu A, Raile K, et al. Fathers of children with diabetes mellitus and their role in coping strategies in the family. J Pediatr Endocrinol Metab 2001;14 Suppl 1:639–43.

28. Al-Agha MA, Majdi WM, Aljefri HM, Ali MA, Al-Agha AE, Abd-Elhameed IA, et al. Effect of parents’ educational level and occupational status on child glycemic control. J Pat Care 2017;3:1000130.

29. Kim SH, Lee SY, Kim CW, Suh YJ, Hong S, Ahn SH, et al. Impact of socioeconomic status on health behaviors, metabolic control, and chronic complications in type 2 diabetes mellitus. Diabetes Metab J 2018;42:380–93.

30. French SA, Wall M, Mitchell NR. Household income differences in food sources and food items purchased. Int J Behav Nutr Phys Act 2010;7:77.

31. Krishnan S, Coogan PF, Boggs DA, Rosenberg L, Palmer JR. Consumption of restaurant foods and incidence of type 2 diabetes in African American women. Am J Clin Nutr 2010;91:465–71.

32. Sørensen KK, Nielsen EP, Møller AL, Andersen MP, Møller FT, Melbye M, et al. Food purchases in households with and without diabetes based on consumer purchase data. Prim Care Diabetes 2022;16:574–80.

33. Affan ET, Praveen D, Chow CK, Neal BC. Comparability of HbA1c and lipids measured with dried blood spot versus venous samples: a systematic review and meta-analysis. BMC Clin Pathol 2014;14:21.

34. Elliott TG, Dooley KC, Zhang M, Campbell HSD, Thompson DJS. Comparison of glycated hemoglobin results based on at-home and in-lab dried blood spot sampling to routine venous blood sampling in-lab in adult patients with type 1 or type 2 diabetes. Can J Diabetes 2018;42:426–32.e1.