Comparison Of Body Mass Index (BMI), an old modality for prediction of gestational diabetes, with the new formulas such as Body Adipose Index (BAİ), Adipose Percentage and Type a Body Fat Index (ABSI): a prospective cohort study

Abstract

Background

GDM increases maternal and fetal complication risks. Fat distribution may predict risk better than total adiposity. While BMI is common, indices like BAI, ABSI, and adiposity percentage may add predictive value. This study compares these indices for GDM prediction and tests whether BMI remains the strongest predictor.

Aim

To compare body fat indices in predicting gestational diabetes mellitus (GDM).

Materials and methods

This prospective study was conducted at a tertiary hospital between March 2024 and November 2024. Pregnant patients presenting during the first trimester were included. Clinical and demographic data, waist and hip circumferences, height, weight, high-density lipoprotein, and triglyceride levels were recorded for each participant. Body mass index (BMI), body adiposity index (BAI), a-type body shape index (ABSI), and adiposity percentage were calculated. Between the 24th and 28th gestational weeks, a 75-gram oral glucose tolerance test was administered. Based on the results, the patients were grouped as GDM-positive or GDM-negative.

Results

Clinical and demographic characteristics, as well as height measurements, were similar between the two groups. However, there were statistically significant differences in relation to weight, hip circumference, waist circumference, BMI, BAI, ABSI, and adiposity percentage (p < 0.005). The ROC analysis revealed that the optimal cut-off value for BMI in predicting GDM was 25.5, with 78% sensitivity and 60% specificity. For BAI, the optimal cut-off value was 30.45, with 72% sensitivity and 64% specificity. The optimal cut-off value for ABSI was 0.069 with61% sensitivity and 48% specificity. Adiposity percentage had an optimal cut-off value of 14.47, with 72% sensitivity and 74% specificity.

Conclusion

BMI, BAI, ABSI, and adiposity percentage are effective formulas for predicting GDM. Among these, BMI remains the most effective predictor.

Introduction

Reported in 2020, the worldwide prevalence of Gestational Diabetes mellitus varies between 2% and 38%, while in Turkey it is 17.6% [1, 2]. During pregnancy, particularly in the third trimester, diabetogenic hormones such as growth hormone, corticotropin-releasing hormone, placental lactogen (chorionic somatomammotropin), prolactin, and progesterone are secreted from the placenta to support fetal growth [3].Pregnancy is characterized by enhanced beta-cell function alongside insulin resistance. GDM occurs in pregnancies where beta-cell dysfunction fails to adequately compensate for the insulin resistance [3].

It is now widely recognized that in metabolic syndrome, which includes insulin resistance, hyperinsulinemia, dyslipidemia, obesity, diabetes mellitus, and hypertension, a central pattern of body fat distribution plays a more significant role than general or regional obesity [4, 5]. While studies in the literature have examined fat distribution in the prediction of type 2 diabetes mellitus and hypertension, there is no research specifically focusing on GDM.

In a cohort study involving many ethnic groups, the duration of pregnancy was found to be associated with body mass index (BMI), but no association was found with other body fat indices [6]. In another study, no association was found between pregnancy loss and BMI, but an association was found with other fat indices [7].

Body composition parameters in early pregnancy (such as neck circumference, hip circumference, waist-to-hip ratio, and visceral fat thickness) increase the likelihood of GDM and can be used for cost-effective early screening [8]. Although recent studies have indicated that ABSI may be superior to BMI in predicting postpartum mortality and cardiovascular risks, its limited advantage in predicting GDM has been emphasized [9].

This study was conducted given the association of GDM with adverse obstetric outcomes such as macrosomic infants, gestational hypertension, polyhydramnios, and stillbirth [10], as well as its high global prevalence. We did not find any studies in the literature comparing body fat indices in GDM.

The aim of this research was to compare body mass index (BMI), a long-established parameter for body fat assessment, with newer formulas such as the body adiposity index (BAI), adiposity percentage, and a-type body shape index (ABSI) in predicting GDM, establish a cut-off value, and provide pre-pregnancy counseling to prevent complications.

Materials and methods

Study population

This prospective, single-center study was conducted at a tertiary care hospital. Between March 2024 and November 2024, all pregnant patients who presented to the Obstetrics and Gynecology Department of Ankara City Hospital during their first trimester, regardless of body structure. The study was conducted using a consecutive patient inclusion method, and randomization was not applied. The design and sampling strategy have been clearly added to the Methods section. Approval was obtained from the Ankara City Hospital Ethics Committee for this study (TABED2 -24-183). Informed consent to participate was obtained from all participants in the study. The principles of the Declaration of Helsinki were adhered to at every stage of the study.

This study included patients diagnosed with GDM who underwent a 75-gram OGTT after an 8-hour fast at 24–28 weeks of gestation [11]. Diagnosis was based on the IADPSG criteria [12].

Gestational age was determined based on the last menstrual period or the crown-rump length measured in the first trimester. The clinical and demographic data of all participants were recorded. Waist and hip circumferences were measured using a tape measure and documented. Height and weight measurements were taken, and BMI, BAI, ABSI, and body adiposity percentage were calculated based on the recorded data.

The hip circumference was measured with a tape measure at the widest point of the gluteal region and recorded. The waist circumference was measured with a tape measure between the navel and the ribs and recorded.

Each patient’s history of gestational diabetes, polycystic ovary syndrome, family history of diabetes, and insulin resistance was assessed and recorded.

BMI was calculated by dividing weight by the square of height in metric units (BMI = weight/height²) [13]. BAI was measured using an automated calculator developed by Bergman et al., which accounts for ethnicity and age (https://webfce.com/bri-calculator/). Metric units were also used for BAI. Adiposity percentage was calculated using the formula, 0.93 × BAI ˗ 14.89 [14].

ABSI was calculated using the following formula developed by Krakauer et al.: ABSI = waist circumference / (BMI^2/3 × height^1/2) [15].

All body measurements, fasting plasma glucose, High-density lipoprotein (HDL)-cholesterol, triglyceride concentrations were also taken during the first trimester.

Excluded from the study were multiple pregnancies, organ transplant recipients, patients with immunodeficiencies, hypertensive or diabetic patients, and individuals with active or chronic viral hepatitis, autoimmune hepatitis, or incomplete/inaccessible data.

Statistical analysis

The sample size was calculated using G*Power (version 3.1) for a two-tailed Mann-Whitney U test; a large effect size (Cohen’s d = 0.80) was assumed. This assumption is based on evidence that pre-gestational BMI elevation shows a moderate-to-large association with GDM risk (e.g., adjusted OR = 3.07 [95% CI: 2.35–4.00] in obese women [16]). A conservative large effect size was chosen due to the exploratory nature of the comparison of new indices. α = 0.05, with a power of 95%, required a minimum of 80 participants.

The Kolmogorov-Smirnov and Shapiro-Wilk tests were employed to assess the normality of the data distribution. The Mann-Whitney U test was used to compare non-normally distributed variables. Descriptive analyses were presented as median (minimum–maximum) for non-normally distributed variables. The chi-square test was employed for categorical variables. Receiver operating characteristic (ROC) curve analysis was conducted to determine the cut-off values of BMI, BAI, ABSI, and adiposity percentage for predicting GDM. Univariate regression analysis was undertaken to evaluate the development of GDM. A p-value of < 0.05 was considered statistically significant.

Results

A total of 146 patients were initially enrolled in the study. However, 41 patients were excluded due to not undergoing OGTT between the 24th and 28th weeks of gestation. As a result, 105 patients were included in the final analysis. These patients underwent a 75-g OGTT between the 24th and 28th gestational weeks. Among them, 18 patients (17%) tested positive for OGTT and were categorized as GDM-positive, while the remaining 87 patients (83%) tested negative and were categorized as GDM-negative.

Table 1 presents the clinical and demographic data, along with HDL-cholesterol, triglyceride concentrations, HDL-cholesterol/triglyceride ratios, fasting plasma glucose levels, gestational age at the time of presentation, height, weight, waist circumference, hip circumference, BMI, body adiposity index, adiposity percentage, and ABSI index for the GDM-positive and GDM-negative groups. The GDM-positive group demonstrated statistically significant differences in weight, hip circumference, waist circumference, BMI, BAI, ABSI, and adiposity percentage compared to the GDM-negative group (p < 0.005) (Table 1).

Table 1.

Clinical and demographic characteristics, biochemical values,and anthropometric measurements of the patients

Variables	GDM-positive n = 18	GDM-negative n = 87	p-value
Age	30.5 (21–40)	29 (19–40)	0.083
Gestational week at presentation	10.0 (5.0–13)	11 (5.0–14)	0.269
Gravida	2.0 (1.0–6.0)	2.0 (1.0–6.0)	0.453
Parity	0.5 (0.0–3.0)	0.5 (0.0–3.0)	0.623
Fasting plasma glucose mg/dl	79.5(70-90)	76.5(59-92)	0.278
HDL -cholesterol (mg/dl)	46.5 (34–90)	55.6(32–97)	0.092
Triglyceride concentration (mg/dl)	119 (47–178)	110 (25–287)	0.767
HDL-cholesterol/triglyceride (mg/dl)	0.47 (0.19–1.26)	0.54 (0.13–2.52)	0.459
Height (cm)	161 (150–173)	163 (150–173)	0.981
Weight (kg)	76.5 (59–115)	65 (43–97)	<0.001
BMI (kg/m2)	31.05 (23.1–42.1)	24.6 (16.2–33.8)	<0.001
Waist circumference (cm)	82 (71-108)	78(58–104)	0.011
Hip circumference (cm)	102 (89-131)	96(66-118)	0.001
BAI	32.4 (23.8–48.6)	27.9 (14.6–40.6)	0.003
ABSI	0.071(0.07-0.09)	0.069(0.06-0.08)	0.013
Adiposity percentage	16.28 (8.29–31.35)	12.09 (0.27–23.91)	0.003

GDM gestational diabetes mellitus, HDL high-density lipoprotein, BMI body mass index, BAI body adiposity index, ABSI a-type body shape index. Statistical test: Mann-Whitney U test, Descriptive analyses were presented as median (minimum–maximum). P value < .05 was considered statistically significant

The ROC analysis revealed that the optimal cut-off value for BMI in predicting GDM was 25.5, with 78% sensitivity and 60% specificity (area under the curve [AUC] = 0.803; p < 0.000). For BAI, the optimal cut-off value was 30.45, with 72% sensitivity and 64% specificity (AUC = 0.724; p = 0.003). The optimal cut-off value for ABSI was 0.069 with61% sensitivity and 48% specificity (AUC = 0.686; p = 0.013). Adiposity percentage had an optimal cut-off value of 14.47, with 72% sensitivity and 74% specificity (Fig. 1).

Fig. 1.

ROC analysis of BMI, BAI, ABSI, adiposity percentage for GDM prediction

Univariate and Multivariate regression analysis showed that BMI was the strongest predictor of GDM (Tables 2 and 3).

Table 2.

Univariate Logistic Regression Analysis of Risk Factors for Gestational Diabetes Mellitus

GDM	p-value	OR	95% confidence interval
			Lower	Upper
Maternal Age	0.106	1.092	0.981	1.216
History of gestational diabetes in a previous pregnancy	<0.001	107.500	12.157	950.554
Insulin resistance	0.260	5.059	0.301	84.888
History of PCOS	0.543	1.687	0.312	9.125
Family history of DM	0.002	16.346	2.867	93.186
fasting glucose levels	0.056	1.996	1.010	8.786
BMI	<0.001	1.346	1.171	1.547
BAI	0.001	1.201	1.076	1.342
ABSI	0.001	3.779	2.373	6.016
Adiposity percentage	0.001	1.218	1.082	1.372

ABSI a-type body shape index, BMI body mass index, BAI body adiposity index, GDM gestational diabetes mellitus, OR odds ratio, PCOS Polycystic Ovary Syndrome P value < .05 was considered statistically significant

Table 3.

Multivariate Logistic Regression Analysis of Risk Factors for Gestational Diabetes Mellitus

GDM	p-value	OR	95% confidence interval
			Lower	Upper
Maternal Age	0.049	1.296	1.001	1.679
History of gestational diabetes in a previous pregnancy	0.004	116.193	4.736	2850.478
Insulin resistance	0.843	8.336	0.676	76443
History of PCOS	0.759	0.512	0.007	36.692
Family history of DM	0.418	3.396	0.176	65.504
fasting glucose levels	0.072	1.896	1.210	6.786
BMI	0.007	1.878	1.192	2.957
BAI	0.114	0.754	0.531	1.070
ABSI	0.127	3.779	2.373	6.016
Adiposity percentage	0.236	0.746	0.604	1.248

ABSI a-type body shape index, BMI body mass index, BAI body adiposity index, GDM gestational diabetes mellitus, OR odds ratio, PCOS Polycystic Ovary Syndrome. P value < .05 was considered statistically significant

Discussion

The International Diabetes Federation and the Pregnancy Study Groups estimate the global prevalence of GDM to be approximately 17% [17]. GDM arises in pregnancies where beta-cell dysfunction fails to adequately respond to insulin resistance. It is generally accepted that central fat distribution plays a more significant role than overall fat distribution in insulin resistance. Studies have shown that obesity significantly increases the prevalence and incidence of type 2 diabetes, with visceral adipose tissue being a strong predictor of type 2 diabetes [18–20] Furthermore, visceral fat accumulation has been associated with complications of type 2 diabetes, such as atherosclerosis and diabetic retinopathy [21, 22]. However, to the best of our knowledge, the literature contains no study specifically focusing on GDM in this context.

Given the known adverse effects of GDM on pregnancy, including polyhydramnios, macrosomia, associated obstetric complications, and intrauterine death, as well as the high global prevalence of GDM, the current study was planned. The central role of insulin resistance in GDM and its association with body fat prompted an investigation into the utility of formulas that estimate body fat in predicting GDM, which formed the basis of this study.

Various methods have been developed to assess body fat. The most accurate methods for measuring body fat are dual-energy X-ray absorptiometry (DXA), computed tomography scans, and magnetic resonance imaging, but they are costly and time-consuming, making them impractical for routine clinical use [23]. BMI is routinely employed in clinical practice to estimate body fat [13]. Some authors have argued that individuals with high muscle mass may exhibit elevated BMI levels despite having low body fat, thereby rendering BMI insufficient for accurately assessing body fat percentage. Consequently, they developed a new formula called “BAI” and introduced an associated measure, adiposity percentage, to address this limitation [24]. It has been found that BAI is effective in estimating body fat percentage [25].

The adipose tissue index measured via DXA has been found to correlate with both BMI and BAI [26, 27]. While some studies indicate that BMI is more effective in reflecting body fat percentage, others have demonstrated a proportional relationship between BAI and BMI [14, 28]. In recent years, another adiposity index, ABSI, has gained prominence. This index, based on waist circumference, was designed with the premise that it may better predict obesity-related mortality [19]. There are studies in the literature comparing such anthropometric measurements for predicting type 2 diabetes and hypertension [29, 30].

Our findings indicate that BMI is the strongest predictor of GDM, while the current literature emphasizes the role of BMI in early pregnancy; one study indicated that early pregnancy BMI increases the risk of GDM and that BMI gain is not associated with this risk, confirming that pregestational BMI is key in prevention [31]. Similarly, a meta-analysis highlighted that a pregestational BMI of 30 or higher significantly increases GDM prevalence; it is also associated with factors such as family history and macrosomia [32]. The combined effect of BMI and ABSI on mortality prediction in adults with diabetes was examined in a cohort, noting that ABSI better reflects central obesity but has limited superiority in the specific context of GDM [33]. Although BAI was evaluated among anthropometric indices for predicting type 2 diabetes in a review, direct comparative data on pregnancy complications are limited [16]. In the current study, formulas for BMI, BAI, adiposity percentage, and ABSI, which are used to estimate visceral fat, were applied and compared for their effectiveness in predicting GDM. Cut-off values for each formula were determined. BMI was found to be more effective than the newly developed formulas.

This study was limited by its single-center design and relatively small sample size, which resulted in a low number of GDM-positive cases. Therefore, analyses assessing the ability of these indices to predict GDM-related complications could not be performed.

The lack of early screening as recommended by ACOG may have affected the overall results of this study; due to the sample size and single-center design, sufficient power was not provided to discuss the effect of early screening, particularly in high-risk subgroups. Future studies should implement early screening and examine its effect on the prediction of GDM using indices such as BMI, BAI, ABSI, and adiposity percentage in multivariate models.

In conclusion, BMI, BAI, ABSI, and adiposity percentage are all effective formulas for predicting GDM. Among these, BMI remains the most reliable due to its longstanding use and ease of clinical application.

Authors’ contributions

All authors have accepted responsibility for the entire content of this manuscript and approved its submission. BBO: Conceptualization, Methodology, Drafting the article, Supervision, Visualization. AT: Conceptualization, Investigation, AAB: Drafting the article, SB: Methodology, Visualization. ENT: Investigation, Data curation, Drafting the article. EK: Visualization normal Analysis. ÖK: Data curation EK: Visualization, Writing – original draft.DS: Analysis and interpretation of data, Validation, Visualization, Writing – review & editing.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript, and this research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data supporting this study is available through the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Approval was obtained from the Ankara City Hospital Ethics Committee for this study (TABED2 -24-183). The principles of the Declaration of Helsinki were adhered to at every stage of the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.