Correlation Between Insulin Suppression Test and Quantitative Insulin Sensitivity Check Index in Hypertensive and Normotensive Obese Patients

Insulin resistance plays a central role in the pathophysiology of diabetes and is associated with obesity and other cardiovascular risk factors (1). In the assessment of insulin resistance, several methods have been developed. The “gold standard” hyperinsulinemic-euglycemic clamp (2) and the insulin suppression test (IST) (3) are two established methods to quantify insulin sensitivity in vivo, but neither is easily applied in large populations. Thus, it is of interest to develop simple methods to estimate insulin sensitivity that are useful for large epidemiological studies. A mathematical model derived from the so-called homeostasis model assessment (HOMA) (4) has been described as a simple and reproducible method in clinical practice. Recently, Katz et al. (5) have described a novel quantitative insulin-sensitivity check index (QUICKI) that shows a high correlation with the hyperinsulinemic-euglycemic clamp. In the present report, we studied the correlation among IST, QUICKI, and HOMA in a population of normotensive-obese (NT-OB) and hypertensive-obese (HT-OB) patients in order to determine their accuracy.

We recruited 20 obese (BMI >30 kg/m²) male patients; 12 were NT-OB, and 8 were newly HT-OB and had never been treated before. None had previous history of metabolic disorders or were on medication with effects on insulin sensitivity. As the control group, 10 healthy (BMI <25 kg/m²) age- and sex-matched volunteers were included. To estimate the insulin sensitivity, we first performed an IST. This test acts by suppressing endogenous insulin secretion with a sustained infusion of somatostatin. Simultaneously, exogenous crystalline insulin is infused at a constant rate to achieve a steady state of plasma insulin (SSPI), and then the resultant steady state of plasma glucose (SSPG), in response to a constant glucose infusion is determined, as we have previously described (6). The insulin sensitivity index (ISI) was calculated with the formula ISI (dl · kg^–1 · min^–1) = [glucose infusion rate (mg kg^–1 · min^–1)/SSPG (mg/dl)] × 10³. The SSPG and ISI levels were considered as measures of insulin sensitivity. Furthermore, we calculated the insulin sensitivity for each subject with the QUICKI formula: 1/log[G₀ + I₀], where G₀ and I₀ are fasting glucose (mmol/l) and fasting insulin (μUI/ml), respectively; and with the HOMA approach: HOMA = g × i/22.5, where g is fasting glucose (mmol/l), and i is the fasting insulin (μUI/ml).

The results showed that fasting insulin concentrations were significantly greater in the HT-OB (32.5 ± 10.9 μUI/ml) and NT-OB (24.8 ± 13.7 μUI/ml) patients than in the control subjects (10.6 ± 3.3 μUI/ml; P < 0.0001 and P < 0.005, respectively). In the obese subjects, especially in the HT-OB group, the lipid profile showed a higher degree of abnormalities. In fact, serum triglyceride and total and LDL cholesterol concentrations in the NT-OB and HT-OB groups were higher than those for the control group, whereas HDL cholesterol levels were lower (P < 0.05 and P < 0.0001, respectively). Furthermore, a significant increase of total and LDL cholesterol levels and a decrease of HDL cholesterol levels were also observed in the HT-OB patients compared with the NT-OB patients (P < 0.05). Lastly, the uric acid levels were higher in the HT-OB group (8.5 ± 0.3 mg/dl) than in the NT-OB group (7.4 ± 0.7 mg/dl) and in the control subjects (7.0 ± 0.5 mg/dl; P < 0.05 and P < 0.0001, respectively).

The SSPG values for HT-OB (237.91 ± 41.67 mg/dl) and NT-OB (182.35 ± 49.54 mg/dl) patients were higher than those for control subjects (126.02 ± 20.7 mg/dl; P < 0.001 and P < 0.05, respectively). In the obese patients, ISI levels (25.9 ± 5.1 dl · kg^–1 · min^–1 for HT-OB and 35.68 ± 11.42 dl · kg^–1 · min^–1 for NT-OB) were lower than those of the control subjects (51.11 ± 9.22 dl · kg^–1 · min^–1; P < 0.001 and P < 0.05, respectively). Both ISI and SSPG values showed that HT-OB patients had a greater degree of insulin resistance than NT-OB patients (P < 0.05). The QUICKI values (0.437 ± 0.011 for HT-OB, 0.478 ± 0.045 for NT-OB, and 0.605 ± 0.052 for control subjects) showed a gradation similar to ISI levels. In fact, QUICKI levels were lower for the HT-OB group than for the NT-OB group (P < 0.02) and the control subjects (P < 0.001). The NT-OB patients also showed QUICKI values lower than those of the control group (P < 0.001). The incidence of insulin resistance was determined from the ISI values attained during the IST. All patients with ISI values below the mean of the control group −2 SD were considered insulin resistant. The same criteria was used to estimate insulin resistance from QUICKI values. The analysis showed that 85.8% of the HT-OB patients and 58.5% of the NT-OB patients were insulin resistant. The calculated HOMA values were 8.79 ± 1.62 for HT-OB, 6.22 ± 2.68 for NT-OB, and 2.12 ± 0.63 for control subjects. Similar to the ISI and QUICKI, the HOMA method showed that the HT-OB patients were more insulin resistant than the NT-OB patients (P < 0.05) and the control subjects (P < 0.001). Also, the NT-OB group had higher HOMA values than the control subjects (P < 0.001).

The overall correlation between QUICKI and ISI was very high (r = 0.888; P < 0.001). Similar values were obtained for each individual group (r = 0.86 for HT-OB patients, P < 0.001; r = 0.76 for NT-OB subjects, P < 0.001; and r = 0.79 for control subjects, P < 0.001) (Fig. 1). Next, we compared SSPG with QUICKI and found a large overall negative correlation between them (r = −0.81; P < 0.001), due to the fact that SSPG is an index of insulin resistance and increases when the insulin sensitivity decreases. As expected, the overall correlation between HOMA and QUICKI was very high (r = −0.91; P < 0.001). In contrast, the overall correlation between HOMA and ISI was significantly smaller (r = −0.69) than that between ISI and QUICKI (r = 0.888; P < 0.05).

The results of the present study clearly show that obese patients with hypertension have a higher incidence of insulin resistance (85.8%) than obese patients without hypertension (58.5%), as estimated from SSPG and ISI data. Similar results were obtained using QUICKI: insulin sensitivity was lowest in the HT-OB group, intermediate in the NT-OB group, and highest in the control group. In our patients, we have not found differences in the incidence of insulin resistance assessed by IST or QUICKI. However, there seems to be a gradation in the severity of the insulin resistance present in obese patients, leading finally to the coexistence of hypertension. In fact, it is well known that insulin resistance is present in most hypertensive patients with obesity. Another major cardiovascular risk factor associated with obesity is an abnormal plasma lipid profile (7). We observed that for the obese patients, the lipid profile closely followed the changes in the insulin sensitivity. Lastly, the uric acid levels in the HT-OB patients confirm that hyperuricemia is an inherent component of the metabolic syndrome.

On the other hand, we found that the overall correlation between QUICKI and ISI was very high (r = 0.888; P < 0.001). As expected, in our study, we also found a very high correlation between HOMA and both QUICKI and ISI. Nevertheless, the correlation between QUICKI and ISI was significantly higher than the correlation between HOMA and ISI. Katz et al. (5) have recently reported that the correlation between QUICKI and hyperinsulinemic-euglycemic clamp measurements of insulin sensitivity was significantly better than the correlation between the minimal-model sensitivity index and glucose clamp. These results suggest that QUICKI contains additional independent information about insulin sensitivity that is not captured by the minimal-model approach. Based on the observed correlations between QUICKI and the hyperinsulinemic-euglycemic glucose clamp and between QUICKI and the ISI, QUICKI may be considered as a very good, inexpensive, and simple tool to estimate insulin sensitivity in a large population with, in particular, a clustering of cardiovascular risk factors.