Thrombin generation in morbid obesity: significant reduction after weight loss


Ingrid Pabinger, Division of Haematology and Haemostaseology, Department of Medicine I, Medical University of Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria.
Tel.: +43 1 40 400 4448; fax: +43 1 40 400 4030.


Summary. Background: Patients with morbid obesity (MO; body mass index > 40 kg m−2) suffer from an increased risk of cardiovascular disease, stroke, venous thromboembolism and all-cause mortality. Objectives: Because weight loss by bariatric surgery reduces cardiovascular and all-cause mortality, we hypothesized that the plasmatic clotting system might be involved in cardiovascular risk. Patients/Methods: Thirty-six MO patients [mean age 42 (±13) years; 29 female) were investigated before and 2 years after bariatric surgery. Thrombin generation was measured with a commercially available assay (Technothrombin-TGA,Technoclone). Metabolic parameters and parameters of the hemostatic system, such as tissue factor (TF), TF pathway inhibitor (TFPI), plasminogen activator inhibitor-1 (PAI-1) and prothrombinfragment 1.2 (F1.2), were determined. To investigate associations of changing parameters, deltas were calculated. Results: Metabolic parameters improved with a mean weight loss of 41 (±19) kg. Postoperatively, the lag phase was significantly extended compared with preoperative values [median (25th–75th percentile), 7 (4–12) vs. 12 (7–19) min, P = 0.005]. Peak thrombin decreased after weight loss from 345 (232–455) to 282 (111–388) nm (P = 0.015) and the area under the curve from 3962 (3432–5023) to 3227 (2202–4030) nm thrombin (P < 0.001). TF, PAI-1 and F1.2 significantly decreased after weight loss. Analyses of the deltas showed a significant correlation between peak thrombin and total cholesterol (r = 0.50), triglycerides (r = 0.46) and HbA1c (r = 0.55). Moreover, an inverse correlation was found between insulin resistance and the lag phase (r = −0.46). Conclusion: Thrombin generation, a marker of the overall coagulation potential, decreased significantly with weight reduction. This might, at least in part, explain the decreased risk of cardiovascular disease after bariatric surgery.


Obesity is a major health concern due to its increasing incidence and high morbidity. In particular, the association of excess body weight and type 2 diabetes mellitus (T2DM) or insulin resistance, arterial hypertension and dyslipidemia is well established [1], leading to an increased risk of cardiovasular diseases, such as coronary heart disease (CHD) and ischemic stroke [2–9]. Also, accumulating evidence suggests an association between obesity, the metabolic syndrome and venous thromboembolism (VTE) [10,11].

Interestingly, in many individuals with obesity changes in the hemostatic and fibrinolytic activity are observed that lead to hypercoagulability. For instance, obesity has been associated with elevated levels of tissue factor (TF), coagulation factors (Fs) VII and VIII, von Willebrand factor (VWF) and fibrinogen [12,13]. An increased activation of the coagulation system is reflected by high levels of prothrombin fragment 1 + 2 (F1.2) [12,14], which is released when activated FX cleaves prothrombin to thrombin, and reflects the in-vivo thrombin generation. Furthermore, morbidly obese patients present a hypofibrinolytic state owing to higher levels of the plasminogen activator inhibitor 1 (PAI-1) [15]. In general, such a prothrombotic state may contribute to the development of cardiovascular disease and VTE in patients with morbid obesity.

A very promising tool to detect an individual’s coagulation potential is the measurement of the in-vitro thrombin generation, which describes the potential of a sample to generate thrombin and may quantify the composite effects of multiple parameters of the coagulation system, thus predicting a prothrombotic state [16,17]. Thrombin is a key enzyme in the coagulation process and leads to the conversion of fibrinogen to fibrin, resulting in clot formation. The thrombin activity can be measured in plasma by continuous cleavage of a chromogenic or fluorescent substrate, and can be registered in a thrombin generation curve. From this curve, various parameters can be inferred that describe thrombin activity, including the time until thrombin burst (lag phase), the peak amount of thrombin generation (peak thrombin) and the total amount of thrombin generated [area under the curve (AUC), representing the endogenous thrombin potential (ETP)].

Weight control remains a widely accepted and recommended clinical measure in patients with obesity [1], in order to reduce the risk of obesity-related morbidity. Studying morbidly obese subjects and their marked weight loss after bariatric surgery is an established method of investigating the effects of changes in body weight and their associations with cardiovascular risk factors. Therefore, in this prospective and longitudinal study we investigated the influence of weight loss on thrombin generation in morbidly obese patients before and 2 years after bariatric surgery. In addition, several other coagulation and metabolic parameters were assessed.

Materials and methods

Study design, study participants and surgical procedure

In this prospective longitudinal study, we consecutively included 36 severely obese patients (29 female, seven male) who planned to undergo bariatric surgery for induction of weight loss. Surgery was indicated for patients with a history of repeated failures of conservative non-surgical techniques and with a BMI > 40 kg m−2. These patients underwent laparoscopic Roux-en Y gastric bypass (RYBP) (n = 26), as described by Mason et al. [18], or gastric banding (n = 10), as described by Kuzmak [19]. All patients were operated on by the same two surgeons.

Because patients after gastric bypass operation randomly start to lose weight and some of the patients gain weight again (on account of malnutrition), we included only those who showed continuing weight loss. As a steady state of weight is normally achieved after 2 years, only patients with a follow-up after this period were eligible.

Exclusion criteria for surgery and for the study were symptomatic CHD, a history of VTE, infectious disease, bulimia or other mental eating disorders, heavy alcohol consumption, major psychiatric disease, hepatic or renal failure, Cushing syndrome, thyroid dysfunction, or other major endocrine disorders, as well as age below 18 years. Patients with oral anticoagulation were also excluded from the study, in order to avoid the influence of anticoagulation on thrombin generation.

For purposes of comparison, thrombin generation was measured as well in a control group of 71 age- and gender-matched healthy individuals who had neither a history of arterial or venous thromboembolism nor bleeding tendency. Their mean BMI was 24 (±4) kg m−2. In both groups, preanalytic treatment, as well as measurement of thrombin generation, was performed identically.

All subjects were carefully instructed about the aims of the study, and written informed consent was obtained. Metabolic, blood coagulation and inflammatory parameters were measured before and 2 years after bariatric surgery, if weight loss was achieved.

Changes in dietary intake, exercise habits and medications

Dietary changes after bariatric surgery were similar to those that have been previously reported [20]. Postoperatively, all patients were instructed to ride on a bicycle for at least 20 min a day. Before surgery, three of the patients were taking statins or fibrates, 10 patients received treatment with angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, five patients were taking β-blockers and two patients were taking calcium channel blockers. In total, 13 patients were treated with antihypertensive drugs before surgery. A total of four patients received oral antidiabetic agents and none of the patients needed insulin. After surgery, none of the patients received statins or fibrates any more, four patients were still treated with angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, one patient was taking a β-blocker, and two patients were taking calcium channel blockers. In total, seven patients still needed antihypertensive drugs postoperatively. A total of three patients received oral antidiabetic agents. Postoperatively, none of the patients needed insulin. Two patients were treated with antiplatelet therapy preoperatively as well as postoperatively.

Laboratory methods

For detection of diabetes, a standard venous oral glucose tolerance test (OGTT) with 75 g glucose was performed, and glucose tolerance status was defined in accordance with the Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus [21].

All blood specimens for the evaluation of laboratory parameters were collected by an atraumatic venipuncture after overnight fasting for at least 10 h. The OGTT was performed in the same session. Blood pressure was measured routinely with a mercury sphygmomanometer under resting conditions. Blood glucose, total cholesterol, high-density lipoprotein (HDL) cholesterol and triglycerides were measured by enzymatic in vitro tests (Roche Diagnostics Corp, Indianapolis, IN, USA), and glycosylated hemoglobin (HbA1c) by high-performance liquid chromatography (Diamat; Bio-Rad, Munich, Germany).

For quantification of insulin concentrations an enzyme-linked immunoassay (ELISA) system was used (Enzymuntest Insulin, ES 600; Boehringer, Mannheim, Germany). Insulin resistance was determined from fasting glucose and insulin concentrations from the homeostatic-model assessment – insulin resistance (HOMA-IR) as described previously [22]. Determination of fibrinogen was performed on a STA coagulation analyzer according to the manufacturer’s recommendations (Diagnostica Stago, Asnieres, France).

Blood samples for measurements of coagulation parameters were drawn into plasma vacuum tubes (Vacuette; Greiner-Bio-One, Kremsmuenster, Austria) containing 1/10 volume sodium citrate stock solution at 0.129 mmol L−1. Samples were then centrifuged to obtain platelet-poor plasma and aliquots were stored at −80 °C until testing was performed in series. F1.2 levels were measured by an ELISA (Enzygnost F1.2; Dade-Behring, Marburg, Germany) according to the manufacturer’s instructions. The measurements of tissue factor (TF) and tissue factor pathway inhibitor (TFPI) were performed with ELISA (IMUBIND Tissue Factor ELISA and IMUBIND® Total TFPI ELISA kit; American Diagnostica Inc., Stamford, CT, USA) according to the manufacturer’s instructions. Thrombin generation was determined also in platelet-poor plasma using a commercially available assay kit (Technothrombin TGA; Technoclone, Vienna, Austria) on a fully automated, computer-controlled microplate reader (FLx800; BioTek, Winooski, VT, USA) and specially adapted software (Technothrombin TGA) using the fluorogenic substrate Z-Gly-Gly-Arg-AMC (Bachem, Bubendorf, Switzerland). In this assay, thrombin generation is initiated in citrated plasma by 71.6 pm recombinant human tissue factor lipidated in 3.2 μm phospholipid micells [phosphatidylcholine (2.56 μm) and phosphatidylserine (0.64 μm)]. According to the manufacturer’s information the final concentration of tissue factor in the reagent used (RClow, Technoclone, Vienna, Austria) is 5 pm. For analysis, the lag phase, the maximum concentration of thrombin (peak thrombin) generated and the AUC were used. Baseline and postoperative samples of all patients were assayed in the same batch to minimize inter-assay variability.

Statistical analyses

Outcome parameters were tested for normal distribution using the Kolmogorov–Smirnov test. Within- and between-group differences were analyzed by Student’s paired and unpaired t-tests. Correlations were calculated using the Pearson correlation. For statistical analyses spss Software Package 15.0 (Chicago, IL, USA) was used. SPSS was used in the standard settings; that is, outliers were included in the analysis, whereas extremes were not. In case of left-skewed distribution (lag phase, F1.2, PAI-1, TF), data were Lg10 transformed, before evaluation with paired standard Student’s t-test was calculated. A P-value < 0.05 was considered as the level of significance. Data are presented as mean ± standard deviation (SD) or median (25th–75th percentile). In order to analyse associations between parameter changes during massive weight loss, deltas were calculated by subtraction of postoperative values from preoperative ones. Pearson’s test was used for all univariate associations apart from delta lag phase time, which was not distributed normally after Lg10 transformation. To investigate the associations of the delta lag phase time, Spearman’s rho test was used. Associations of thrombin generation parameters with baseline parameters were subjected to multivariate analysis, in order to further elucidate which changes in baseline parameters can be linked to changes in thrombin generation.


Baseline patient characteristics and effect of weight loss on metabolic parameters

Thirty-six severely obese patients (29 female, seven male) were included in our study. All of these patients underwent bariatric surgery for weight loss. The mean age (±standard deviation, SD) of the study participants at study inclusion was 42 (±13) years. Patients were investigated before and 2 years after surgery. Parameters of thrombin generation were compared with a healthy, age- and gender-matched control group with a BMI of 24 (±4) kg m−2.

Bariatric surgery induced a mean weight loss of 41 (±19) kg (pre- and postoperative BMI: 45 (±5) vs. 31 (±6) kg m−2; P < 0.001). Clinical characteristics of the study population including metabolic parameters at baseline and postoperatively are given in Table 1. After weight loss the following parameters had improved significantly: systolic blood pressure, HDL-cholesterol, triglycerides, fasting insulin and insulin concentrations after 120 min, glucose concentrations after 120 min, HOMA-IR and HbA1c. In addition, the systemic inflammation marker CRP (mg dL−1) was significantly decreased after weight loss [1.3 (±0.9) vs. 0.4 (±0.3), P = 0.001].

Table 1.   Patient characteristics and metabolic parameters before and after bariatric surgery
  1. Data are mean ± SD. P-values given are paired standard student’s t-test. P-values < 0.05 were considered statistically significant. BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; HOMA, Homeostasis-Model Assessment; HbA1c, glycolysated hemoglobin; CRP, C-reactive protein.

Weight (kg)130 ± 2789 ± 2341 ± 19< 0.001
BMI (kg m−2)45 ± 531 ± 614 ± 6< 0.001
Blood pressure (mmHg)
 Systolic148 ± 19127 ± 1721 ± 16< 0.001
 Diastolic90 ± 1383 ± 97 ± 150.053
Total cholesterol (mg dL−1)189 ± 40180 ± 349 ± 430.366
LDL-cholesterol (mg dL−1)105 ± 3392 ± 2813 ± 300.071
HDL-cholesterol (mg dL−1)56 ± 1962 ± 18−6 ± 120.047
Triglyceride (mg dL−1)139 ± 79101 ± 5138 ± 550.008
Insulin (μU mL−1)
 Insulin fasting19 ± 98 ± 511 ± 8< 0.001
 Insulin 120 min75 ± 4616 ± 2259 ± 560.001
Glucose (mg dL−1)
 Glucose fasting86 ± 1579 ± 217 ± 190.122
 Glucose 120 min106 ± 2472 ± 2634 ± 320.001
HOMA4.2 ± 2.61.5 ± 0.82.8 ± 2.20.001
HbA1c (%)6.1 ± 1.25.5 ± 0.70.6 ± 1.10.028
CRP (mg dL−1)1.3 ± 0.90.4 ± 0.30.9 ± 0.90.001

Effect of weight loss on thrombin generation and procoagulant activity in morbidly obese patients

Weight loss was associated with a statistically significant alteration of parameters of thrombin generation, such as lag phase, peak thrombin and the AUC (Table 2 and Fig. 1). Postoperatively the lag phase was significantly prolonged compared with preoperative values [median (25th–75th percentile), 7 (4–12) vs. 12 (7–19) min, P = 0.005]. Peak thrombin had decreased after weight loss from 345 (232–455) to 282 (111–388) nm (P = 0.015) and the AUC from 3962 (3432–5023) to 3227 (2202–4030) nm thrombin (P = < 0.001) (always pre- vs. postoperatively). Preoperatively, these parameters of thrombin generation generally deviated clearly from those measured in our 71 healthy, age- and gender-matched control group members [lag phase, 14 (12–17) vs. 7 (4–12) min, P < 0.001; peak thrombin, 175 (143–261) vs. 345 (232–455) nm, P < 0.001; AUC, 3413 (3045–3750) vs. 3962 (3432–5023) nm thrombin, P = 0.001] (always control group vs. preoperative patient values). Postoperative values reached almost the levels of the control group [lag phase, 14 (12–17) vs. 12 (7–19) min, P = 0.74; AUC, 3413 (3045–3750) vs. 3227 (2202–4030) nm thrombin, P = 0.07] (always control group vs. postoperative patient values). However, peak thrombin was still significantly higher in morbidly obese patients postoperatively compared with our normal-weight control group [175 (143–261) vs. 282 (111–388) nm, P = 0.04]. Moreover, other parameters of the hemostatic system, such as circulating TF, PAI-1 and F1.2 concentrations, decreased significantly after weight loss (Table 2). However, no significant changes in concentrations of TFPI and fibrinogen were observed after weight loss.

Table 2.   Thrombin generation and other hemostatic parameters before and after bariatric surgery
  1. AUC, area under the curve; F1.2, prothrombin fragment 1.2; PAI-1, plasminogen activator inhibitor-1; TF, tissue factor; TFPI, tissue factor pathway inhibitor. Data are median (25–75th percentile). In case of left-skewed distribution, data were Lg10 transformed before evaluation with paired standard student’s t-test. P-values < 0.05 were considered statistically significant. Given P-values reflect the significance of difference between pre- and postoperative.

Thrombin generation
 Lag phase time (min)7 (4–12)12 (7–19)−4 (−8 to 3)0.005
 Peak thrombin (nm)345 (232–455)282 (111–388)56 (−61 to 255)0.015
 AUC (nm thrombin)3962 (3432–5023)3227 (2202–4030)734 (84–1906)< 0.001
F1.F2 (pmol L−1)1200 (867–1200)173 (83–595)907 (2–1048)0.009
PAI-1 (U mL−1)2.8 (1.2–8.2)0.9 (0.7–1.8)1.0 (0–5.1)0.007
TF (pg mL−1)139 (84–196)57 (0–130)57 (6–123)0.025
TFPI (ng mL−1)74 (64–85)71 (55–80)2 (−8 to 15)0.383
Fibrinogen (mg dL−1)409 (325–443)381 (337–427)−22 (−65 to 66)0.973
Figure 1.

 Box plot analysis of the distribution of parameters of thrombin generation (lag phase, peak thrombin and AUC): in patients with morbid obesity pre- vs. postoperatively. The boundaries of the box represent the 25th and the 75th percentiles, respectively. The line inside the box marks the median. Following the SPSS Box-whisker function the 10% and the 90% percentiles were used for whiskers. Open circles represent outliers.

Univariate and multivariate analysis of the associations between parameters of thrombin generation and metabolic parameters

We evaluated the correlation coefficients (univariate) between some key variables of thrombin generation and metabolic parameters pre- and postoperatively and calculated the corresponding coefficients for the deltas. These data are summarized in Table 3. At baseline a significant correlation was observed between peak thrombin and fasting insulin levels (r = 0.63), lag phase and HDL-cholesterol (r = −0.47). Postoperatively, AUC and glucose levels after 120 min (= −0.49) correlated significantly. Furthermore, the deltas of peak thrombin significantly correlated with total cholesterol (r = 0.50), triglycerides (r = 0.46) and HbA1C (r = 0.55). Moreover, an inverse correlation was detected between deltas of lag phase and HOMA-IR (r = −0.46, P = 0.043 ). In a second step, a multivariate stepwise backward regression analysis regarding the delta peak thrombin values was performed including all delta variables studied for association with delta peak thrombin, as depicted in Table 3. Only delta weight was omitted so as not to confound it with delta BMI. Multiple stepwise backward regression revealed that final determinants of delta peak thrombin were delta cholesterol (β = 0.498, P = 0.018) and delta HOMA-IR (β = 0.433, P = 0.035). The multivariate significant association of delta HOMA-IR with delta peak thrombin was not caused by confounding of delta cholesterol because the Betas changed from uni- to multivariate only from 0.431 (P = 0.058) to 0.433 (P = 0.035), and thus by only 0.5%.

Table 3.   Associations of thrombin generation parameters pre- and postsurgery as well as delta
Lag phase time*Peak thrombinAUCLag phase time*Peak thrombinAUCLag phase time*Peak thrombinAUC
  1. BMI, body mass index; HDL, high-density lipoprotein; HOMA-IR, Homeostasis-Model Assessment – insulin resistance; HbA1c, glycosylated hemoglobin. Pearson’s test was used for all associations apart from delta lag phase time, which was not distributed normally after Lg10 transformation. To investigate the associations of delta lag phase time Spearman’s rho test was used. *Logarithmically transformed variables.

Weight (kg)0.410.05−0.150.51−0.100.650.300.220.200.410.220.360.240.490.120.630.270.27
BMI (kg m−2)0.160.490.060.780.230.310.320.190.160.520.320.180.130.770.250.290.380.11
Insulin fasting (μU mL−1)−0.030.900.630.010.310.210.230.37−0.220.380.230.37−0.350.080.360.170.250.36
Glucose 120 min (mg dL−1)−0.200.420.120.630.080.76−0.230.37−0.350.16−0.490.05−0.020.83−0.130.50−0.330.82
HDL-cholesterol (mg dL−1)−0.470.03−0.010.990.140.53−0.170.480.030.910.110.640.270.950.200.410.120.63
Cholesterol (mg dL−1)−0.240.28−0.110.62−0.210.350.370.120.120.620.120.62−0.220.950.500.030.320.19
Triglyceride (mg dL−1)−0.170.450.140.520.160.480.420.070.220.380.160.52−0.230.450.460.050.300.22
HbA1c (%)−0.340.12−0.000.99−0.020.93−−


In this study we demonstrated a considerable decrease in thrombin generation in morbidly obese patients after weight loss induced by bariatric surgery. We observed that plasma obtained from patients preoperatively showed an earlier and higher thrombin generation, as shown in the lag phase, peak thrombin and AUC, compared with postoperative values and matched controls. Moreover, we measured other parameters of the hemostatic system, such as TF, F1.2 and PAI-1, which were also significantly decreased after weight loss. Our data suggest that an enhanced prothrombotic state and signs of coagulation activation are present in patients with morbid obesity and can be improved after weight loss. As expected, these changes in hypercoagulability and coagulation activation were accompanied by an improvement of metabolic and hemodynamic parameters after weight loss.

The increased risk of cardiovascular morbidity and mortality in patients with morbid obesity [1,2,4,5] is reduced after bariatric surgery, and likewise a reduction in traditional cardiovascular risk factors can be observed [6,8]. In particular, we found that several biomarkers were related to this increased risk, such as markers of inflammation (e.g. CRP) [7], dyslipidemia, hypertension or glucose intolerance. However, the classical risk factors cannot explain why up to 20% of all coronary events occur in total absence of these risk factors [23]. This leads us to assume that hemostatic disturbances in terms of hypercoagulability might play an important role in this context.

Interestingly, recent studies have reported an association of increased thrombin generation with hyperglycemia [24], coronary artery disease [25] and VTE [16,17]. Cimenti et al. [26] described significantly higher thrombin generation levels in severely obese children, when compared with age-matched, normal-weight, healthy controls. Up to now, data on thrombin generation in adult obese patients and the effects of weight loss on thrombin generation have not been available. Our results show an increased thrombin generation in patients with morbid obesity and may thus provide an additional explanation for the high cardiovascular risk observed in these patients. Interestingly, as the risk of the individual patient is reduced by weight loss [8], our patients contemporaneously show a reduction of thrombin generation.

The relevance of hemostatic disturbances in our morbidly obese patients is supported by the changes in the levels of F1.2, a marker of in-vivo thrombin generation that is indicative of ongoing thrombin activation, and TF, the main initiator of coagulation activation that could lead to increased thrombin generation. In addition, we were able to confirm previous results regarding elevated TF [12], PAI-1 [15] and CRP [7] levels and their reduction after weight loss. Preoperatively as well as postoperatively, most parameters of thrombin generation were not sufficiently correlated with metabolic markers, such as dyslipidemia and insulin resistance or BMI. Thus, the latter variables could not help to elucidate the association of thrombin generation with morbid obesity and weight loss. Likewise, Sola et al. [14] did not find a significant association of F1.2 and metabolic markers after weight loss.

Therefore, in our study, the changes from pre- to postoperative values of hemostatic and metabolic parameters, as well as the respective correlation coefficients, were calculated. Accordingly, deltas of thrombin generation and deltas of metabolic parameters were investigated for mutual associations: analyses of the deltas showed a significant correlation between peak thrombin (mainly dependent on phospholipids/microparticles contained in the sample, as well as on the protein C pathway and antithrombin) and parameters of dyslipidemia, such as cholesterol and triglycerides, and HbA1c. A decrease in lipid concentrations and HbA1c levels was associated with a decrease in peak thrombin. Moreover, an inverse correlation was found between HOMA-IR and the lag phase (mainly dependent on tissue factor added as trigger and TFPI contained in the sample), which may indicate an association between insulin resistance and increased thrombin generation. This possibility is underpinned by our findings in the multivariate analysis, in which delta cholesterol and delta HOMA-IR accounted for about 43.5% of the variation in the major target parameter of our study, peak thrombin generation.

Some limitations, but also some strengths of our study need to be addressed. First of all, the small sample size poses a limitation of our study, although it allowed us to detect significant changes in thrombin generation before and after weight loss, also in comparison to a control group. Because we performed sequential measurements of other coagulation markers in our study, we could assess additional changes of parameters of hypercoagulability and coagulation activation, which also decreased after weight loss induced by bariatric surgery. However, further studies are needed to confirm the impact of thrombin generation on the cardiovascular risk in morbidly obese patients. Furthermore, different methods for measurement of thrombin generation are available. In some studies the use of corn trypsin inhibitor was discussed to block contact pathway activation. In our study we did not use corn trypsin inhibitor. However, according to van Veen et al. [27], for tissue factor concentrations from 5 pm upwards there is no effect of corn trypsin inhibitor on the endogenous thrombin potential/thrombin generation. Furthermore, ongoing discussion indicates that the methods we applied in our study for the measurement of thrombin generation should not be used to measure the endogenous thrombin potential, if it is not corrected for α2-macroglobulin. According to Chandler et al. [28], a correction for α2-macroglobulin is not necessary for low sample volumes as we used in our study.

In conclusion, our findings confirm the close relationship between morbid obesity and alterations in the coagulation system. Bariatric surgery for weight loss, which has been proven to significantly improve life expectancy and reduce the cardiovascular risk in morbidly obese patients, leads to a significant reduction in thrombin generation, the key process in hemostasis. Because thrombin generation significantly decreases after weight loss, this decrease may contribute to a reduction in the cardiovascular risk generally associated with morbid obesity.


All the authors were involved in drafting the article or revising it critically for important intellectual content, and all the authors have approved the final version to be submitted for publication. L. Ay and G.-H. Schernthaner had full access to all study data and hold responsibility for the integrity of the data and the accuracy of data analysis. Study conception and design: L. Ay, H.-P. Kopp, G.-H. Schernthaner, I. Pabinger, G. Schernthaner. Acquisition of data: L. Ay, J.-M. Brix, C. Ay, P. Quehenberger, G.-H. Schernthaner. Analysis and interpretation of data: L. Ay, H.-P. Kopp, J.-M. Brix, C. Ay, P. Quehenberger, G.-H. Schernthaner, I. Pabinger, G. Schernthaner. Writing of manuscript: L. Ay, G.-H. Schernthaner, C. Ay, I. Pabinger.


We would like to thank S. Koder for determining the parameters of thrombin generation. Furthermore, we thank T. Altreiter for proof-reading this manuscript and Technoclone for their assistance.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.