Sagittal abdominal diameter and visceral adiposity: correlates of beta-cell function and dysglycemia in severely obese women

Abstract

Background: In the context of increasing obesity prevalence, the relationship between large visceral adipose tissue (VAT) volumes and type 2 diabetes mellitus (T2DM) is unclear. In a clinical sample of severely obese women (mean body mass index [BMI], 46 kg/m(2)) with fasting normoglycemia (n = 40) or dysglycemia (impaired fasting glucose + diabetes; n = 20), we sought to determine the usefulness of anthropometric correlates of VAT and associations with dysglycemia.

Methods: VAT volume was estimated using multi-slice computer tomography; anthropometric surrogates included sagittal abdominal diameter (SAD), waist circumference (WC) and BMI. Insulin sensitivity (Si), and beta-cell dysfunction, measured by insulin secretion (AIRg) and the disposition index (DI), were determined by frequently sampled intravenous glucose tolerance test.

Results: Compared to fasting normoglycemic women, individuals with dysglycemia had greater VAT (P < 0.001) and SAD (P = 0.04), but BMI, total adiposity and Si were similar. VAT was inversely associated with AIRg and DI after controlling for ancestry, Si, and total adiposity (standardized beta, -0.32 and -0.34, both P < 0.05). In addition, SAD (beta = 0.41, P = 0.02) was found to be a better estimate of VAT volume than WC (beta = 0.32, P = 0.08) after controlling for covariates. Receiver operating characteristic analysis showed that VAT volume, followed by SAD, outperformed WC and BMI in identifying dysglycemic participants.

Conclusions: Increasing VAT is associated with beta-cell dysfunction and dysglycemia in very obese women. In the presence of severe obesity, SAD is a simple surrogate of VAT, and an indicator of glucose dysregulation.

Conflict of interest statement

Conflict of interest No conflict of interest

Figures

Fig. 1

Receiver operating characteristic curves of VAT, SAD, WC and BMI for prediction of dysglycemia in obese women. The reference line depicts a probability of 50 %, i.e., random chance. ‡P<0.001, compared to the reference line; *P=0.039, compared to the reference line. The optimal cut-points for each parameter (with their respective sensitivity and specificity) for distinguishing between normo- and dysglycemia appear are as follows: VAT, cut-off=5,059 cm3, sensitivity=65.0 %, specificity=85.0 %; SAD, cut-off 30.7 cm, sensitivity= 55.0 %, specificity=80.0 %; WC, cut-off 132.7 cm, sensitivity= 60.0 %, specificity=60.5 %; BMI, cut-off=49.8 kg/m2, sensitivity= 35.0 %, specificity=80.0 %