Carbohydrate metabolism in obese women: association with the circulating levels of adiponectin and retinol binding protein-4
José Silva-Nunes1,2, Alice Melão2, Ana Oliveira2, Léone Duarte1, Miguel Brito2 & Luísa Veiga2
Background: Adiponectin is an adipokine with anti-diabetic properties. Recently, retinol binding protein-4 (RBP4) was identified as an adipokine associated with insulin resistance (IR) and a potential marker for type 2 diabetes (T2D) development.
Aims: To compare adiponectin and RBP4 circulating levels according to presence/absence of diabetes/prediabetes, in obese women. To study the degree of association of those adipokines with indexes of IR and insulin secretion (IS).
Methods: This study was conducted after approval by Hospitals Ethical Committee. We studied 84 obese women anthropometrically characterized. They were submitted to fasting blood sample collection for adiponectin and RBP4 assessments, followed by an oral glucose tolerance test. We used two IS indexes (homeostatic model (HOMA%beta) and insulinogenic index (INS-i)), three insulin sensitivity indexes (Matsuda formula (MATSUDA), McAuley index (McAULEY) and QUICKI) and one IR index (HOMA-IR).
Results: Patients were characterized by mean age=34.1±8.2 years, BMI=43.6±8 kg/m2, waist circumference=117.7±15.3 cm, waist:hip ratio (WHR)=0.88±0.08, adiponectin=6.72±3.33 μg/ml, RBP4=30.57±13.92μg/ml, HOMA%beta=261.8±175.8, INS-i=24±20.1, MATSUDA=3.93±2.92, McAULEY=6.59±1.62, QUICKI=0.14±0.01 and HOMA-IR=4.42±3.67. Normoglycemia was present in 57 patients. Adiponectin and RBP4 were inversely correlated (P=0.01; r=−0.347). Normoglycemic women exhibited higher adiponectin levels (P=0.035), QUICKI (P<0.001), McAULEY (P=0.007), MATSUDA (P=0.006) and INS-i (P=0.013) and lower HOMA-IR (P<0.001); they also presented non-significantly (P=0.078) lower RBP4. Adiponectin, but not RBP4, was inversely associated with WHR (P=0.002; r=−0.332). Adiponectin was negatively correlated with HOMA-IR (P=0.006; r=−0.298) and HOMA%beta (P=0.035; r=−0.232) and directly with QUICKI (P<0.001; r=0.455), MATSUDA (P<0.001; r=0.472) and McAULEY (P<0.001; r=0.465). RBP4 was directly associated with HOMA-IR (P=0.028; r=0.265) and inversely with QUICKI (P=0.035; r=−0.254), MATSUDA (P=0.023; r=−0.273) and McAULEY (P=0.037; r−=0.251).
Conclusions: The IR in diabetic/prediabetic women accompanies a decrease in adiponectin and increase in RBP4 levels. Adiponectin presents a parallel inverse association with IR and anthropometry. Contrarily, RBP4 associates with IR in a non-anthropometrically dependent way. This suggests a RBP4 specific pathway in compromising insulin sensitivity.