Improvement of Insulin Resistance and Early Atherosclerosis in Patients after Gastric Banding
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Abstract
Objective: To evaluate the effect of massive weight loss on insulin sensitivity, soluble adhesion molecules, and markers of the insulin resistance syndrome (IRS).
Research Methods and Procedures: Eighteen morbidly obese patients underwent gastric banding and were evaluated before and 6 and 12 months after surgery. Total insulin secretion, hepatic insulin extraction, and insulin sensitivity were analyzed by oral glucose-tolerance test model analysis. In addition, soluble intercellular adhesion molecule-1, vascular cell adhesion molecule-1, E-selectin, leptin, high-sensitivity C-reactive protein, plasminogen activating factor-1 (PAI-1), and tissue plasminogen activator were measured.
Results: BMI dropped from 45.22 ± 5.62 to 36.99 ± 4.34 kg/m2 after 6 months and 33.72 ± 5.55 kg/m2 after 12 months (both p < 0.0001). This intervention resulted in a significant reduction of blood pressure (p < 0.00001), triglycerides (p < 0.01), fasting blood glucose (p = 0.03), basal insulin (p < 0.001), and basal C-peptide (p = 0.008) levels. Total insulin secretion decreased (p < 0.05), whereas hepatic insulin extraction (p < 0.05) and oral glucose insulin sensitivity index (p < 0.0001) increased compared with baseline. Leptin (p < 0.0001) and E-selectin levels decreased significantly after 6 and 12 months (p = 0.05), whereas significantly lower levels of intercellular adhesion molecule-1 and PAI-1 were only seen after 6 months. Subclinical inflammation, measured by high-sensitivity C-reactive protein, was lowered to normal ranges. No changes were observed in vascular cell adhesion molecule-1 and tissue plasminogen activator levels.
Discussion: Although gastric banding ameliorates several features of the IRS, including 29.05% improvement in insulin sensitivity and blood pressure and reduction of soluble adhesion molecules and PAI-1, considerable weight loss did not normalize all components of the IRS in morbidly obese patients.
Introduction
Obesity is associated with the insulin resistance syndrome (IRS),1 which is characterized by hyperinsulinemia, dyslipidemia, hypertension, and coagulation abnormalities (1, 2). The IRS is related to cardiovascular disease (CVD) and associated with an increased morbidity and mortality attributable to premature atherosclerosis (3, 4, 5, 6, 7, 8). Insulin resistance (IR) is suggested to be accompanied with subclinical chronic inflammation, which might be a contributor to the development of atherosclerosis. C-reactive protein is a prototypic marker of inflammation, and numerous studies have confirmed that detection of high-sensitivity C-reactive protein (hsCRP) can be used to predict CVD (9, 10, 11).
Other components of the IRS are adhesion molecules (AMs), which are reported to be early indices of endothelial activation (12, 13, 14, 15). AMs regulate the interaction between leukocytes and endothelium, such as leukocyte adhesion to the vessel wall, endothelial infiltration by mononuclear cells, and the proliferation of smooth cells. These alterations are not only key steps in the response to tissue injury but also to inflammation.
Therefore, subclinical chronic inflammation and the up-regulation of circulating AMs (cAMs), together with the alteration of platelet function, could be the earliest steps toward premature atherosclerosis; in addition, these indicators have been found to correlate with higher levels of glucose, insulin, and leptin (16, 17). However, the role of cAMs in the IRS is still uncertain.
Hence, because any feature of the IRS strongly correlates with increasing BMI, morbidly obese patients (BMI >40 kg/m2) are at highest risk for CVD and, therefore, should be encouraged to loose weight (18). Laparoscopic adjustable gastric banding (LAGB) is a safe and proven method for achieving long-term weight loss and major improvements in comorbidities in severe obesity (19, 20, 21).
The purpose of this study was to investigate the effect of considerable weight loss on markers of the metabolic syndrome and IR in morbidly obese subjects. In addition, we evaluated whether the reduction of markers of early atherosclerosis was related to weight loss or IR.
Research Methods and Procedures
Patients
Between January 1999 and November 1999, 18 morbidly obese patients (15 women and 3 men; mean age, 42.56 ± 9.74 years) attending our nutrition department were investigated before and 6 and 12 months after LAGB. Patients who fulfilled the criteria for surgical treatment according to the American Bariatric Association (22) and who had a BMI ≥40 kg/m2 were included in the study and underwent a LAGB (Lap-Band; Inamed, Santa Barbara, CA) at the Surgical Department of the Sozialmedizinisches Zentrum Ost in Vienna. During the study period, no severe postoperative complications were observed in any patient. One-year follow-up for all subjects was achieved, with interim evaluation performed quarterly. Routine blood checks to exclude acute or chronic infectious diseases were taken, and dietary advice was given at these visits. The nutrition management that we used has also been previously described (23). The day after operation, patients were instructed to follow a liquid diet for 4 weeks. Thereafter, a solid low-fat diet was given, and a list of rules developed for patients with gastric banding to avoid vomiting was provided. However, the energy intake varied among individuals, depending on the compliance and on the band calibration. Band adjustments were performed only in the case of weight stabilization (<4 kg in the last months) and a low vomiting frequency. At each visit, patients had their body weight (to the nearest 0.1 kg) and height (to the nearest 0.1 cm) recorded while wearing light indoor clothes and no shoes. For identifying upper-body obesity, waist was measured at the level of the umbilicus while the patient was in a supine position.
At each visit, systolic and diastolic blood pressures were measured twice, with the subject in a supine position after 5 minutes of rest; the measurement was taken on the right arm using an appropriately sized cuff, with a width of at least 40% of the circumference of the arm. The mean of the two measurements was used to determine blood pressure.
In all patients, an upper-body ultrasound examination was performed preoperatively for identifying steatosis hepatis or cholecystolithiasis. All patients gave written informed consent after the local Ethical Committee had approved the study protocol.
Study
At baseline and 6 and 12 months after LAGB, patients were tested for impaired glucose tolerance [oral glucose-tolerance test (OGTT), 75 grams] according World Health Organization (WHO) criteria. The study group represented a typical morbidly obese population, and patients on antidiabetic medication and/or insulin therapy were excluded.
Impaired glucose tolerance was diagnosed when the 2-hour plasma glucose exceeded 7.8 mM (140 mg/dL); diabetes was diagnosed when the 2-hour plasma glucose was >11.1 mM (200 mg/dL). An OGTT was performed on each subject at baseline and at 6 months, but only 14 OGTTs were available at 12 months because four patients refused to have a third OGTT. In these cases, only routine blood checks were performed. None of the patients was on any antidiabetic medication or insulin therapy. Drugs that could influence insulin sensitivity (e.g., ACE- and ATII-antagonists and statins) were stopped 2 days before the OGTT.
The standard dose of 75 grams of glucose in H2O solution was administered. In the morning after an overnight fast, venous blood samples were collected for insulin, C-peptide, and glucose measurements in the fasting state and at 10, 20, 30, 40, 60, 90, 120, 150, and 180 minutes after glucose load.
In addition, in the fasting condition, lipids and liver function parameters, leptin, the cAMs intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), E-selectin, plasminogen activating inhibitor-1 (PAI-1), and tissue plasminogen activator (t-PA) were measured.
Lipids, liver function parameters, and the acute-phase hsCRP, white blood count, and fibrinogen were measured according to standard laboratory techniques. Blood was rapidly centrifuged, and plasma and serum aliquots were stored −72 °C until further analysis. Plasma glucose values were assessed with the hexokinase methods applied on a modula analyzer (Roche, Basel, Switzerland). Insulin (Serono Diagnostics, Freiburg, Germany), C-peptide (CIS Bio International, Cedex, France), leptin (Human Leptin RIA kit, Linco Research, Inc., St. Charles, Mo.) were tested by commercially available radioimmunoassay kits with an interassay coefficient of variation of <5% for insulin and C-peptide and 4.1% for leptin.
Serum samples for determination of cAMs were stored at −72 °C and analyzed by commercially available ELISAs (British Bio-technology Product Ltd., Abington, United Kingdom), as described previously (24, 25), with both inter- and intra-assay coefficients of variation being <6%. t-PA antigen (normal range, 1 to 12 ng/mL) was measured by ELISA COALIZA t-PA (Chromogenix AB, Mölndal, Sweden); free plasminogen activator inhibitor 1 antigen (4 to 49 ng/mL) was determined by an ELISA method from Technoclone (Vienna, Austria).
Data Analysis
Insulin Secretion and Appearance
OGTT data were analyzed by a two-compartment mathematical model (26) that reconstructs the patterns per unit volume of C-peptide secretion and post-hepatic insulin appearance into peripheral circulation. This model yields a figure of the β-cell function and of insulin degradation in the liver. From the concomitant analysis of C-peptide and insulin data, the model allows the estimation of the basal fasting pre-hepatic insulin secretion rate per unit volume, basal secretion rate (BSR) (measured in picomoles per liter per minute), of the total amount of secreted insulin per unit volume, total insulin secretion (TIS) (measured in nanomoles per liter per 3 hours), and of the hepatic insulin extraction (HIE) (as the percentage of the secreted hormone) during the entire test (i.e., not only the first pass).
Insulin Sensitivity
From a comprehensive mathematical model, a formula of glucose clearance was derived (27) to obtain an OGTT-based index of oral glucose insulin sensitivity index (OGIS) (measured in milliliters per minute per square meters). OGIS has been validated against the corresponding value obtained during the “gold standard” glucose clamp (27) that has already been used successfully in other studies (28).
The ability to dispose of glucose that relates insulin action to the prevailing insulin amount is given by the disposition index (measured in nanomoles per meters squared), calculated as the product OGIS x ΔAUCins (29), where ΔAUCins is the suprabasal component of AUCins, being the areas under the concentration curve (AUCs) calculated with the trapezoidal rule. The capacity of the β cell to adapt to changes in insulin sensitivity is given by the adaptation index (measured in nanomoles per minute square meters), calculated as the product OGIS × ΔTIS (30), where ΔTIS is the suprabasal component of TIS. Both indexes provide a quantitative figure of the overall metabolic status by simultaneously accounting for insulin action and secretion.
Statistical Analysis
Results are expressed as means ± SD. The paired Student's t test was used for group comparison groups before and after LAGB. To evaluate correlations between cAMs, BMI, insulin sensitivity, and hsCRP, we calculated Kendall's τ correlation coefficients. Statistical analyses were calculated with Origin 5.0 release for Windows (OriginLab, Northampton, MA). A two-sided p < 0.05 was considered statistically significant.
Results
Clinical Characteristics
Body weight, BMI, waist, glucose, insulin, C-peptide, and leptin serum levels are shown in Table 1. Subjects lost a mean of 23.27 kg 6 months after and 33.24 kg 12 months after LAGB, and their BMI dropped from 45.22 ± 5.62 kg/m2 at baseline to 36.99 ± 4.34 kg/m2 6 months after and 33.72 ± 5.55 kg/m2 12 months after LAGB (both p < 0.0001). Waist circumference was significantly reduced after 6 and 12 months (both p < 0.0001).
| Pre-LAGB | 6 Months after LAGB | 12 Months after LAGB | |
|---|---|---|---|
| Patients (n) | 18 | 14 | |
| Age (years) | 42.56 ± 9.74 | ||
| Body weight (kg) | 128.56 ± 16.84 | 105.29 ± 14.4* | 93.71 ± 15.84¶ |
| BMI (kg/m2) | 45.22 ± 5.62 | 36.99 ± 4.34* | 33.72 ± 5.55¶ |
| Waist (cm) | 131.53 ± 15.04 | 120.18 ± 11.5* | 110.82 ± 11.74*,¶ |
| Systolic blood pressure (mm Hg) | 149.38 ± 19.23 | 138.5 ± 16.34* | 132.58 ± 15.32* |
| Diastolic blood pressure (mm Hg) | 92.12 ± 12.71 | 85.64 ± 10.83* | 79.03 ± 9.6* |
| Total cholesterol (mg/dL) | 216.56 ± 21.87 | 209.38 ± 26.15 | 204.88 ± 28.42 |
| High-density lipoprotein (mg/dL) | 49.69 ± 8.98 | 52.46 ± 7.62 | 53.19 ± 7.62 |
| Low-density lipoprotein (mg/dL) | 131.87 ± 22.16 | 128.62 ± 28.73 | 123.38 ± 33.48 |
| Triglycerides (mg/dL) | 162.69 ± 40.76 | 143.82 ± 46.22‡ | 121.00 ± 51.90‖ |
| GPT (mg/dL) | 20.22 ± 2.18 | 12.35 ± 6.51† | 9.11 ± 3.63† |
| YGT (mg/dL) | 24.17 ± 26.53 | 18.25 ± 13.92† | 12.0 ± 3.68† |
| Blood glucose (mg/dL) | 123.11 ± 62.49 | 102.78 ± 37.92† | 100.57 ± 27.25§ |
| Insulin (μU/mL) | 26.59 ± 11.68 | 14.77 ± 6.61‡ | 16.53 ± 9.88‖ |
| C-peptide (ng/mL) | 403.67 ± 94.67 | 339.22 ± 95.6‡ | 274.64 ± 80.44*,‖ |
| Leptin (ng/mL) | 27.75 ± 8.99 | 18.94 ± 9.59* | 13.0 ± 5.38†,¶ |
- Data are means ± SD.
- * p < 0.0001 before vs. 6 months and 6 months vs. 12 months.
- ‡ p < 0.01 before vs. 6 months.
- † p < 0.05 before vs. 6 months and 6 months vs. 12 months.
- § p < 0.05 before vs. 12 months.
- ‖ p < 0.01 before vs. 12 months.
- ¶ p < 0.0001 before vs. 12 months.
Blood pressure was lowered significantly by 9.06 ± 0.44 (6.14%) and 17.64 ± 3.74 (10.15%) mm Hg systole and 6.48 ± 1.88 (7.04%) and 13.09 ± 2.85 (14.21%) mm Hg diastole 6 and 12 months after LAGB, respectively (all p = 0.00001). In the lipid profile, a significant improvement was observed only in triglycerides (p < 0.01), although low-density lipoprotein-cholesterol was lowered by 6.44%, and high-density lipoprotein-cholesterol increased by 7.04% after 1 year. Liver function parameters (glutaimatpyruvat transaminase and y-glutamyl transpeptidase) were both significantly reduced after LAGB (both p < 0.05 after 6 and 12 months LAGB)
Fasting blood glucose (p = 0.03), basal insulin (p < 0.0001), and basal C-peptide (p = 0.008) decreased significantly after 6 months. After 12 months, basal glucose and insulin concentration remained essentially stable, whereas C-peptide showed a further significant reduction after comparing results 6 and 12 months (p = 0.05). Leptin concentration decreased significantly after 6 and 12 months (p = 0.0002 and p = 0.03, respectively).
OGTT and Insulin Sensitivity
According to the WHO criteria, at the baseline OGTT and before LAGB, six patients were diabetic, four were impaired glucose tolerant (IGT), and eight were normal glucose tolerant (NGT). After 6 months, all IGT subjects, except one, and one diabetes subject converted to NGT, and two diabetes subjects converted to IGT; no new case of diabetes was diagnosed. After 12 months, OGTT results were available for 14 of 18 patients. All patients who were NGT, except one, who converted to diabetic, remained NGT. All patients diagnosed as IGT and diabetic at 6 months demonstrated the same WHO classification 12 months after LAGB, except two diabetic patients who converted to IGT (Table 2).
| Patient | Pre-LAGB (n = 18) | 6 Months after LAGB (n =18) | 12 Months after LAGB (n = 14) |
|---|---|---|---|
| 1 | NGT | NGT | NGT |
| 2 | NGT | NGT | NGT |
| 3 | NGT | NGT | NGT |
| 4 | IGT | NGT | DM |
| 5 | IGT | NGT | NGT |
| 6 | DM | IGT | IGT |
| 7 | NGT | NGT | NGT |
| 8 | DM | IGT | IGT |
| 9 | DM | DM | IGT |
| 10 | NGT | NGT | NGT |
| 11 | NGT | NGT | NGT |
| 12 | DM | DM | IGT |
| 13 | NGT | NGT | NGT |
| 14 | NGT | NGT | NGT |
| 15 | DM | NGT | NA |
| 16 | DM | DM | NA |
| 17 | IGT | IGT | NA |
| 18 | IGT | NGT | NA |
- DM, diabetes mellitus; NA, not available.
Metabolic parameters derived from the OGTT analysis are presented in Table 3. After 6 months, insulin secretion decreased both in terms of BSR (p = 0.008) and total amount (TIS, p = 0.05), whereas HIE (p < 0.0001) increased. In addition, BSR and TIS levels were significantly lower after 12 months in comparison with levels before surgery (p = 0.0003 and p = 0.02) and after 6 months (p = 0.05 and p = 0.02). In HIE, no further change was observed. The OGTT-derived insulin sensitivity (OGIS) improved at 6 and 12 months follow-up (both p < 0.0001). No significant changes were observed in the disposition and adaptation indexes at both follow-up visits, although the disposition index tended to decrease after surgery (p = 0.06).
| Pre-LAGB | 6 Months post-LAGB | 12 months post-LAG | |
|---|---|---|---|
| Basal secretion rate (pM/min) | 82.18 ± 19.27 | 69.06 ± 19.45‡ | 55.93 ± 16.09†,‖ |
| Total insulin secretion (nM/3 hours) | 30.44 ± 9.96 | 27.26 ± 6.42† | 21.14 ± 7.23†,§ |
| Hepatic insulin extraction (%) | 63.36 ± 11.81 | 74.23 ± 8.07* | 73.76 ± 10.18§ |
| Insulin sensitivity (mL/min/m2) | 343.0 ± 86.4 | 405.11 ± 86.53* | 442.64 ± 94.77¶ |
| Disposition Index (nmol/m2) | 22.77 ± 3.69 | 17.77 ± 2.72 | 15.74 ± 11.0 |
| Adaption Index (nmol/min/m2) | 5.55 ± 0.79 | 6.08 ± 0.53 | 4.77 ± 2.71 |
- Data are means ± SD.
- * p < 0.0001 before vs. 6 months.
- ‡ p < 0.01 before vs. 6 months.
- † p < 0.05 before vs. 6 months and 6 months vs. 12 months.
- § p < 0.05 before vs. 12 months.
- ‖ p < 0.01 before vs. 12 months.
- ¶ p < 0.0001 before vs. 12 months.
Markers of Inflammation
All markers showed a decrease; however, these results did not show a statistical significance when comparing the results before LAGB and after the surgical procedure (Table 4).
| Before | 6 Months after LAGB | 12 Months after LAGB | |
|---|---|---|---|
| High-sensitivity C-reactive protein (mg/dL) | 1.18 ± 1.2 | 0.77 ± 0.73 | 0.95 ± 1.29 |
| White blood cell count (per mm3) | 8.62 ± 1.76 | 8.26 ± 2.19 | 7.49 ± 2.17 |
| Fibrinogen (mg/dL) | 3.94 ± 0.91 | 3.73 ± 0.74 | 3.47 ± 0.81 |
- Data are means ± SD. All not significant.
AMs and Thrombogenic Risk Factors
Results are summarized in Table 5. Highly significant decreases of soluble ICAM-1 (316.52 ± 129.47 ng/mL before and 257.59 ± 80.18 ng/mL 6 months after; p = 0.03) and E-selectin levels (60.07 ± 30.36 ng/mL before and 41.54 ± 13.76 ng/mL after 6 months; p < 0.001) were observed after 6 months. No statistically significant difference in ICAM-1 levels was seen after 12 months compared with baseline (301.47 ± 98.14 ng/mL), whereas E-selectin concentration was significantly lower (43.75 ± 20.44 ng/mL; p = 0.05) after 12 months compared with levels before surgery.
| Pre-LAGB | 6 Months after LAGB | 12 Months after LAGB | |
|---|---|---|---|
| ICAM-1 (ng/mL) | 316.52 ± 129.47 | 257.59 ± 80.18† | 301.47 ± 98.14 |
| E-selectin (ng/mL) | 60.07 ± 30.36 | 41.54 ± 13.76‡ | 43.75 ± 20.44§ |
| VCAM-1 (ng/mL) | 662.70 ± 108.08 | 675.93 ± 133.08 | 627.94 ± 59.73 |
| PAI-1 (ng/mL) | 57.46 ± 56.63 | 39.3 ± 22.51† | 42.92 ± 17.26 |
| t-PA (ng/mL) | 6.67 ± 2.48 | 5.53 ± 0.97 | 6.85 ± 2.85 |
- Data are means ± SD.
- ‡ p < 0.001 before vs. 6 months.
- † p < 0.05 before vs. 6 months.
- § p < 0.05 before vs. 12 months.
VCAM-1 concentrations did not differ significantly before, 6 and 12 months after LAGB. A positive correlation between ICAM-1 and fasting insulin levels were seen over the study period (r = 0.34; p = 0.05 at baseline, r = 0.36; p = 0.04 and r = 0.41; p = 0.04 at 12 months). E-selectin was positively correlated with BMI before (r = 0.39, p = 0.02) and after 12 months (r = 0.4, p = 0.04). Both AMs did not correlate with OGIS or markers of inflammation.
Analysis of time-dependent interactions revealed that changes in ICAM-1 (ΔICAM-1) after 6 months correlated positively with changes in fasting insulin levels (Δinsulin: r = 0.38; p = 0.04) after 6 months and after 12 months (r = 0.40; p = 0.04). ΔICAM-1 tended to correlate with changes in BMI (ΔBMI: r = −0.38, p = 0.06) and correlated significantly with changes in body weight (Δkg: r = 0.41; p = 0.04) 12 months after LAGB. A positive correlation between changes in E-selectin (ΔE-selectin) and changes in BMI (ΔBMI: r = −0.38, p = 0.05) and changes in fasting insulin (Δinsulin: r = 0.43; p = 0.03) were observed at 12 months, but not at 6 months follow-up.
In addition, PAI-1 decreased significantly (p = 0.03) after 6, but not after 12, months. Levels of t-PA and fibrinogen did not change significantly after LAGB.
Discussion
Our results clearly confirm that large weight loss after LAGB resulted in a significant improvement of several features of the IRS, including an amelioration of glucose hemostasis in the morbidly obese patients. Furthermore, our findings provide evidence that (surgically-induced) weight loss correlates with cAMs.
Obesity-associated comorbidities are reported to be improved or even reversed by severe weight loss after LAGB (19, 20, 21). In this study, a weight loss of 27.11% of the initial body weight was followed by a significant reduction of blood pressure and lipids and lowering of liver function parameters, which might represent reduction of nonalcoholic steatohepatitis (31).
Results of the present study confirmed that severe weight loss after LAGB induced an increase in insulin sensitivity, expressed as OGIS (32). Although clamp technique is the gold standard, recent studies have demonstrated that OGIS gives similar results when differences among groups are tested and the relationships between OGIS and other metabolic variables are studied (27, 28).
In addition to the significant reduction of hyperinsulinemia and insulin hypersecretion, HIE increased significantly 6 months after LAGB, indicating that the reduction of the prevailing insulin levels during an OGTT are not only attributable to a decreased glucose-stimulated insulin secretion, but also to an increased metabolic clearance rate of the hormone, which might be a result of the reduction of nonalcoholic steatohepatitis. In contrast, the decrease of C-peptide levels over the study period (insulin levels were comparable at 6 and 12 months) reflected simply a further decrease in insulin secretion, whereas insulin extraction remained unchanged after 12 months. Consequently, the significant, but not marked improvement, of OGIS by ∼18.11% 6 months after and 29.05% 12 months after LAGB was not followed by a significant change in the compensatory mechanisms that regulate the balance between IR and secretion. In fact, only a trend toward a reduction of the disposition index was observed after LAGB, which indicates that the degree of weight loss of our population was not yet able to fully normalize glucose tolerance. This phenomenon might be attributable to the fact that our patients remained still morbidly obese (BMI = 33.17 kg/m2 after 12 months) and insulin resistant. Another possibility could be the relatively short study period of 12 months, whereas a longer duration of LAGB might have led to normal glucose homeostasis (33).
Disorder in glucose hemostasis may induce an elevated expression of soluble forms of AMs, which might reflect an early stage of atherosclerosis that might be lowered by glycemic control in diabetic patients (34). In the current study, even though the sample size was small and many markers showed a considerable interindividual variability, we found a statistically significant decrease of cAMs. ICAM-1 and E-selectin levels were significantly lower (−16.8% and −25%) after 6 months, whereas after 12 months, only E-selectin levels were significantly reduced, but both sAMs concentrations were still lower compared with baseline levels. VCAM-1 levels were similar in all groups. This was in accordance with other studies (14, 35) reporting elevated VCAM-1 levels only in patients with clinically relevant coronary heart disease. Whereas E-selectin levels showed a positive correlation with BMI, ICAM-1 correlated positively with fasting insulin levels before and after LAGB. Interestingly, time-dependent changes of circulating E-selectin and ICAM-1 concentrations were correlated with changes in BMI and fasting insulin, indicating that obesity and hyperinsulinemia/IR might contribute directly to endothelium activation, followed by an increase of cAMs. Previous studies were conducted in only moderately obese patients, with a maximum weight loss of 10% after multidisciplinary programs, and information about glucose tolerance status was mostly absent. Ferri et al. (36) reported a significant reduction of ICAM-1, E-selectin, and VCAM-1 after a 12-week caloric restriction, a 9% weight loss in obese hypertensive nondiabetic patients, and a correlation with BMI. Recently, others (37) found a significant reduction of soluble ICAM-1 and E-selectin and a correlation of both markers with obesity in obese women after a 3-month lifestyle-modification.
Ziccardi et al. (38) demonstrated increased levels of ICAM-1, VCAM-1, and proinflammatory cytokines in obese women (BMI = 37.2 kg/m2) who significantly decreased their initial body weight by 10% after 1 year.
Furthermore, measures of subclinical chronic inflammation, such as hsCRP or leptin, might have contributed to endothelial activation (11, 14, 15, 16, 17), although we did not observe a correlation with any cAMs in this study. Inflammatory markers decreased to normal ranges after the study period, and no statistical significance was observed in hsCRP because of low baseline levels in this population. The largest decrease was seen in patients with elevated levels over the normal range before surgery and 6 months after LAGB, and these changes were paralleled by the decrease of PAI-1 levels.
Our findings give further support to the hypothesis that, after the start of dieting, weight loss results in the rapid reduction endothelial-activation markers (36, 37, 38). In our study, surgically induced, considerable weight loss resulted in the continuation of the reduction of markers 1 year after LAGB. Because weight loss in obesity is generally associated with changes of the IRS found in this study, including lowering of blood pressure and leptin and lipid levels, we were not able to exclude these features as having beneficial effects on endothelial activation as well (39). Thus, these data provide the possibility that considerable weight loss after LAGB alone does not correct all mechanisms by which morbid obesity impacts endothelial function.
In summary, the remarkable reduction of weight, resulting in decreases in IR and in serum levels of many cardiovascular risk factors, might be useful in ameliorating the pro-atherosclerotic milieu associated with morbid obesity. However, whereas E-selectin and markers of chronic subclinical inflammation showed a continuous decrease over the time course of our study, the effect of LAGB on other AMs and thrombogenic factors seemed to be attenuated.
Acknowledgment
There was no outside funding/support for this study. We would like to express our sincere thanks to Dr. Martha M. Eibl from the Immunological Outpatient Department (Vienna) for offering pertinent comments and Heinz Tüchler (Ludwig Boltzmann Institute of Leukemia, Vienna) for his advice.
