CAN ROTATIONAL THROMBOELASTOMETRY PREDICT THROMBOTIC COMPLICATIONS IN RECONSTRUCTIVE MICROSURGERY? JONAS KOLBENSCHLAG, M.D.,1* ADRIEN DAIGELER, M.D.,1 SARAH LAUER, M.D.,2 GERHARD WITTENBERG, M.D.,3 SEBASTIAN FISCHER, M.D.,2 NICOLAI KAPALSCHINSKI, M.D.,1 MARCUS LEHNHARDT, M.D.,1 and OLE GOERTZ, M.D.1
Thrombotic occlusion of the microvascular pedicle is the major reason for flap loss. Thus, identifying patients who are at risk for such events is paramount. Rotational thromboelastometry (RTE) is widely used to detect coagulopathy and hypercoagulable states. The aim of our study was to assess its diagnostic value in reconstructive microsurgery. In all 181 patients undergoing free tissue transfer at our department between February 2010 and November 2011 preoperative RTE was performed. In addition, coagulation values as well as patient’s demographic data, cause and localization of defect, type of flap and surgical revisions were recorded. The majority of patients was male (59.6%) with traumatic (59.7%) defects located on the lower extremity (60.3%). ALT was the most often used flap (35.9%). Preoperatively, 36.5% of patients had a hypercoagulable RTE (higher than physiological RTE values; intrinsic (ICPT) or extrinsic (ECPT) mean clot firmness (MCF) >72mm or functional fibrinogen (ICF) MCF >25mm). A total of 28 primary thrombosis of the microvascular pedicle occurred, 11 of those in-patients with a hypercoagulable state. Total flap loss rate because ofthrombosis was 7.7% (n 5 14). Both a hypercoagulable RTE assay and a functional fibrinogen to platelet ratio (FPR) of >43 (MCF value of ICF divided by the MCF value of ICPT) were significant predictors of thrombotic flap loss when performing multivariate binary logistic regression, co-factoring for age, sex, and comorbidities (p 5 0.036 and 0.003, respectively). RTE seems to be able to identify patients that are prone to thrombotic complicaC 2013 Wiley Periodicals, Inc. Microsurgery 34:253–260, 2014. tions and might be used as a screening tool. V
Free tissue transfer has had a significant impact on reconstructive surgery over the past decades. The techniques for microvascular tissue transfer have steadily evolved, resulting in high success rates.1 But even despite this tremendous progress, flap loss remains a looming threat with devastating consequences for both patients and surgeons. Flap loss most often occurs because of thrombosis of the microvascular pedicle and may be secondary to local and technical factors (e.g. intima damage, vessel kinking, technical errors regarding the anastomosis) or systemic influences (e.g. acquired or hereditary coagulation disorders like a posttraumatic hypercoagulable state or resistance to activated protein C).2–4 Therefore, the identification of patients with such risk factors and the modification of coagulation to prevent pedicle thrombosis has been an important topic since the early days of microsurgery.5 But even today, the perioperative management of anticoagulation in microvascular tissue transfer remains a highly controversial topic.6 Recent surveys show that a wide range of different anti1 BG University Clinic Bergmannsheil, Department of Plastic Surgery, Burn Center, Bochum, Germany 2 BG Trauma Center, Department of Plastic and Hand Surgery, Ludwigshafen, Germany 3 BG Trauma Center, Department of Anaesthesiology and Intensive Care Medicine, Ludwigshafen, Germany *Correspondence to: Jonas Kolbenschlag, M.D.; BG University Clinic Berg€rkle-de-la-Campmannsheil, Department of Plastic Surgery, Burn Center, Bu Platz 1, 44789 Bochum, Germany. Email: [email protected] Received 14 May 2013; Revision accepted 3 October 2013; Accepted 4 October 2013 Published online 20 October 2013 in Wiley Online Library (wileyonlinelibrary. com). DOI: 10.1002/micr.22199
Ó 2013 Wiley Periodicals, Inc.
coagulatory regimes is in place today.7 Such anticoagulatory therapy is seldom tailored to the specific patients risk profile, mostly because of the difficulties in identifying such patients preoperatively. While an immense amount of manpower, time, and technical utilities is dedicated to flap monitoring in the early postoperative period, less attention has been paid to the preoperative identification of potential high-risk patients. Although such close monitoring is essential for early recognition of pedicle thrombosis and potential flap salvage, a preoperative screening of the coagulatory state of these patients might allow for a targeted anticoagulatory therapy, aid in reducing such incidents and could therefore potentially improve the success rates of free tissue transfer even further. This is especially true as coagulatory disorders are quite common. Factor-V-Leiden Thrombophilia for example is prevalent in up to 15% of the European population.8,9 Other pro-thrombotic conditions such as sickle cell trait even occur in up to 1 in 3 persons in selected populations.10 As factor analysis and in-depth coagulation diagnostics are too costly to perform in every patient undergoing free tissue transfer, a global screening test for coagulatory disorders is needed. Rotational thromboelastometry (RTE) has first been described by Hartert in 1948.11 It generates a real-time image of in vitro clot formation and therefore takes into account not only the plasmatic components of coagulation but also its cellular elements and their interactions. By summing up these factors, it can act as a global screening test of the coagulatory state. RTE has seen an increasing use in various fields because of its point-of-care-
254
Kolbenschlag et al.
availability and real-time results. It is now widely used to detect coagulopathy and hypercoagulable states in perioperative and trauma care and has even been shown to predict myocardial infarction. Recently it was also used as a diagnostic tool in severe burns.12–18 However, very little data exist about the diagnostic value of RTE in microvascular tissue transfer. To our knowledge, only a small preliminary study dealing with this subjected has been published up to now. In 29 patients undergoing free tissue transfer for head and neck reconstruction, Parker et al. found that certain RTE values might be able to predict thrombotic complications in microvascular surgery.19 Therefore, the aim of our study was to assess the diagnostic value of RTE regarding thrombotic complications in reconstructive microsurgery.
ter. Maximal clot firmness (MCF) was defined as the maximum diameter of the clot, measured in millimeters. A hypercoagulabe state leads to increased activation of coagulation, resulting in a larger thrombus and therefore an increased MCF. The physiological range of MCF for ECPT and ICPT is 50–72 and 9–25 mm for ICF as supplied by the manufacturer. The physiological ranges for the other values are also given in Figure 1. In this graphic display of results, time was plotted against mm of clot thickness, resulting in an image of the actual thrombus and its behavior over time. All three activated tests for extrinsic and intrinsic coagulation pathway of thrombus formation and the isolated contribution of functional fibrinogen to the clot are displayed below each other in Figure 1, with the graphic display on the left and numeric results on the right side.
PATIENTS AND METHODS
Hypercoagulable RTE and Functional Fibrinogen to Platelet Ratio
A retrospective review of all patients undergoing free tissue transfer at our department between February 2010 and November 2011 was conducted. In all of the 181 consecutive patients included, pre-operative RTE was R -Device (Tem Internaperformed using the ROTEMV tional GmbH, Munich, Germany) as a standard of care. Rotational Thromboelastometry
RTE is based on an in vitro clotting process. It generates a real-time image of in vitro clot formation and therefore takes into account not only the plasmatic components of coagulation but also its cellular elements and their interactions. By summing up these factors, it can act as a global screening test of the coagulatory state. For RTE-tests, blood samples were given into a cup, which housed a rotating pin. Under continuous circulation of the sample, the blood began to clot. Although it is possible to measure native thrombus formation without any additional activation, activated tests offer improved reliability and reproducibility. For such tests, by activation of either the extrinsic coagulation pathway thrombus formation (ECPT) or intrinsic coagulation pathway thrombus formation (ICPT), thrombus formation was initiated by adding tissue factor for the extrinsic and partial thromboplastin for the intrinsic pathway. In addition, by stimulating the extrinsic pathway and adding a platelet-inhibiting reagent (Cytochalacin), the isolated contribution of functional fibrinogen to the clot was detected (ICF; isolated contribution of fibrinogen). After activation, the in vitro clot formation began and was continuously measured for all three tests to generate a real-time image of several clotting attributes. Coagulation time (CT) was defined as the duration between the initiation of clot formation and the first visible clotting. The time after which the clot achieved a diameter of 20 mm was referred to as clot-formation-time (CFT). The a-angle characterized the kinetics with which the clot gained firmness and diameMicrosurgery DOI 10.1002/micr
In the study presented here, a complete RTE, as performed for each patient, consisted of those three activated tests. On the preoperative day, blood was drawn for standard laboratory work-up, including a coagulation screening and RTE measurements. RTE was considered hypercoagulable when the MCF for the ECPT and ICPT was higher than 72 mm, or higher than 25 mm for ICF. If the RTE values were within the physiological range as supplied by the manufacturer (Fig. 1), RTE was considered physiological. For calculating the functional fibrinogen to platelet ratio (FPR) the MCF of ICF was divided by the MCF of ICPT, resulting in a numeric ratio. For the last analyzed value we chose to use a cutoff value within the physiological range (66 mm) for ECPT-MCF (Mean Firmness of the clot in the extrinsic pathway test). In addition preoperative fibrinogen-, thrombocyte-, hemoglobin-levels as well as PTT and AntiThrombin III activity (AT III) were recorded. We also recorded patient’s sex, age, comorbidities, cause and localization of defect, type of flap used and surgical revisions with an emphasis on thromboembolic events. No pre-operative anticoagulation was given other than the previous medication, postoperative anticoagulation treatment was administered at discretion of the surgeon, either as continuous infusion of unfractionated heparin (15.000U/24h) or as low-molecular heparin subR 0.4 ml, two times a day). cutaneously (ClexaneV Statistical analysis of the data was performed utilizing SPPS 21 for Mac (IBM Corporation) and Excel 2011 for Mac (Microsoft Corporation). For comparison between groups, student’s t-test was used, whereas the correlation of influencing variables with thromboembolic complications and flap loss was calculated using binary logistic regression. Additionally, the odds ratio was calculated for the mentioned risk factors.
ROTEM in Microsurgery
255
R ), ICPT (INTEMV R ), and ICF Figure 1. Hypercoagulable RTE in a patient with thrombotic flap loss. From top to bottom, ECPT (EXTEMV R ) were depicted. The graph represents the actual in vivo clot formation. The clot’s diameter (isolated contribution of fibrinogen; FIBTEMV in mm (MCF) was plotted against time in minutes. Note the elevated MCF in all three tests. Also, the physiological range for each value is given in brackets (CT, clotting time; CFT, clot formation time; ML, maximum lysis; A10 / A20, clot diameter at 10 / 20 minutes). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
The area under the curve (AUC) was calculated to display the predictive potential of certain RTE values. Results are displayed as mean 6 SD. A significance level of P < 0.05 was assumed.
RESULTS
A total of 181 patients with free flaps were included, of which 108 (59.6%) were male. Mean age at the time of operation was 50.2 614.1 years. Thirty-seven per cent of patients were active smokers, 35% had a history of hypertension, 17% were obese (BMI > 25), 9% suffered from peripheral arterial disease (PAD), 8% from diabetes. A history of thromboembolic events was present in 4% of all patients. Soft-tissue defects necessitating free-flap coverage were most often because of trauma (59.7%) and malignancies (24.9%). Most defects were located on the lower extremity (60.3%) followed by the trunk (21.5%), upper extremity (16.6%), and head/neck (1.6%; Table 1). The most often used type of flap was the antero-lateral thigh flap (ALT) in 35.9%, followed by the
Table 1. Distribution of Comorbidities, Cause of Defect and Location of Defect Among the Included Patients Comorbidity/Cause of Defect/Localization Smoking Hypertension Obesity (BMI >25) Diabetes History of thrombosis Peripheral Arterial Disease (PAD) Trauma Malignancy Infection Burns Chronic Ulcer Lower extremity Trunk Upper extremity Head/Neck
No. of patients
%
67 64 31 14 7 16
37% 35.3% 17% 7.7% 3.8% 8.8%
108 45 12 9 7 109 39 30 3
59.7% 24.9% 6.6% 4.9% 3.9% 60.3% 21.5% 16.6% 1.6%
latissimus dorsi flap (24.9%), deep inferior epigastric artery perforator flap (DIEP; 16.5%), parascapular flap (11.6), free gracilis flap (4.5%), fibula flap (3.9%), lateral Microsurgery DOI 10.1002/micr
256
Kolbenschlag et al.
arm flap (1.6%), and superficial gluteal artery perforator flap (S-GAP; 1.1%; Table 2). Mean duration of hospital stay was 46.8 638.69 days (Median 34 days, Range 9–252 days). In 66 patients (36.5%) the pre-operative RTE was deemed hypercoagulable (ECPT or ICPT MCF >72 mm or ICF MCF >25 mm), whereas in the remaining patients the values were within standard range. In all but one patient, if one value (ECPT, ICPT, ICF) was within pathological range the others were too. In this patient, ECPT and ICF were elevated while ICPT was borderline normal with 72 mm. A hypocoagulable state was not seen. The age distribution between patients with pathological and physiological RTE did not significantly differ. Patients with a hypercoagulable RTE were much more likely to be men (P 5 0.0003) with traumatic defects (P 5 0.0076) of the lower extremity (P 5 0.0015). Hypertension, obesity, and history of smoking were not significantly different between the two groups (P 5 0.6, 0.22 and1, respectively). Considering the laboratory measurements, concentrations of fibrinogen and thrombocytes were significantly higher in patients with hypercoagulable RTE (P 5 0.0001 for both). AT III activity was also significantly higher in
Table 2. Distribution of Free Flaps Among the Included Patients Type of flap
No. of patients
%
65 45 30 21 8 7 3 2
35.9% 24.9% 16.5% 11.6% 4.5% 3.9% 1.6% 1.1%
ALT Latissimus dorsi DIEP Parascapular Gracilis Fibula Lateral Arm S-GAP
ALT, anterior lateral thigh flap; DIEP, deep inferior epigastric artery perforator flap; S-GAP, superior gluteal artery perforator flap.
these patients (P 5 0.001), whereas the partial thromboplastin time was longer (P 5 0.019). Hemoglobin concentration on the other hand was higher in patients with a physiological RTE (Table 3). Overall, 28 pedicle thrombosis occurred during the postoperative course of treatment (15.5%). Venous thromboembolisms were found in 15 patients, whereas arterial thrombosis was found in six patients. In seven patients, both arterial and venous vessels were occluded by thromboembolisms. Seventeen of these thrombosis occurred in the 115 patients with physiological RTE (15%), whereas the remaining eleven were found in the 66 patients with a hypercoagulable RTE (16.6%). There was no significant difference in the onset of primary thrombosis between the subgroups (P 5 0.19). The total rate of free flap loss because of thrombotic events was 7.7%, with 14 free flaps being lost to thrombosis. Nine of these flap losses occurred in patients with a pathological RTE (13.6%), whereas the other five were seen in the remaining 115 patients (4.3%). In the group with thrombotic flap losses, fibrinogen levels were significantly higher than in patients without flap losses (P 5 0.024), while there was no such difference regarding thrombocyte level (P 5 0.85; Table 4). In addition, the first thrombosis in flaps which were subsequently lost to thrombotic events occurred a mean of 5.6 days later in patients with hypercoagulable RTE when compared to those with physiological ones (3.6 63.2 vs. 9.2 68 days). Also, the salvage rate in flaps affected by thrombotic complications was considerably lower in the group with pathological RTE findings. While 12 out of the 17 flaps (71%) with thromboembolic complications could be salvaged in patients with physiological RTE, this was only possible in two out of 11 (18%) flaps in the hypercoagulable group (Fig. 2).
Table 3. Comparison of Patients with Pathological and Physiological RTE
Age Sex (male / female) Trauma (yes / no) Malignancy (yes / no) Hypertension (yes / no) Smoking history(yes / no) Lower extremity(yes / no) Obesity (yes / no) Fibrinogen (mg/dl) Thrombocytes (Tsd./ul) PTT (s.) Hemoglobin (g/dl) Antithrombin III (%) PTT, partial thromboplastin time.
Microsurgery DOI 10.1002/micr
Physiological RTE (n 5 115)
Hypercoagulable RTE (n 5 66)
P-value
51.07 613,4 57/58 60/55 40/75 41/74 43/72 59/56 23/92 320.8 610.6 292.35 6104.3 34.1 66.9 12.6 61.8 78.77 614
48.82 615,28 51/15 48/18 6/60 21/45 24/42 50/16 8/58 461.9 694.9 469.52 6175.6 36.47 65.6 10.9 61.7 87.7616.5
0.3 0.0003 0.0076 0.0001 0.6 1 0.0015 0.22 0.0001 0.0001 0.019 0.0001 0.001
ROTEM in Microsurgery
Thirteen out of the 14 thromboembolic flap losses (92%) occurred in the 134 patients with an ECPT-MCF of 66 and higher (9.7% flap loss rate). When calculating the functional FPR using the RTE device (ICF/ICPT ratio, FPR), we found that only 38 patients had a ratio of >43 (21% of all patients). Out of the 14 flap losses, eight (53%) occurred in these patients. All three values (pathological RTE, ECPT-MCF>65, FPR >43) were introduced into multivariate binary logistic regression models adjusting for sex, age, and comorbidities (PAD, history of smoking or thrombosis, diabetes, arterial hypertension, obesity). Considering primary thrombotic events, none of the variables was of significant influence. However, both a pathological RTE and a FPR >43 were strong predictors of thrombotic flap loss (P 5 0.036 and 0.003, respectively; Table 5). None of the other computed values showed a significant correlation with thrombotic flap losses. Table 4. Comparison of Coagulation Values in Patients with and Without Thrombotic Flap Loss
ECPT (50–72mm) ICPT (50–72mm) ICF (9–25mm) Fibrinogen (mg/dl) Thrombocytes (Tsd./ul)
No flaploss (n 5 167)
Flaploss (n 5 14)
P-value
69,9 6 6,14 67,62 6 6,93 23,46 6 10,59 362 6 120 357 6 160.7
73,43 6 7,08 70,93 6 5,42 26,93 6 9,78 441 6 112 349.1 6 138.1
0,074 0,08 0,23 0,024 0.85
ECPT, extrinsic coagulation pathway thrombus formation; ICPT, intrinsic coagulation pathway thrombus formation; ICF, isolated contribution of fibrinogen.
257
DISCUSSION
RTE was first described over 60 years ago, but had not found broad clinical application in the beginning.11 This was mostly because of the complex and interference-prone nature of these early devices. In recent times, RTE devices have become much more user friendly because of automation of the performed test and improved data analysis and handling capabilities. The R device (Tem International GmbH, Munich, ROTEMV Germany) is a point-of-care device capable of performing four parallel measurements, resulting in easy and fast to obtain coagulation measurements. ECPT and ICPT depict activation of the coagulatory cascade either triggered by tissue factor (ECPT) or partial thromboplastin (ICPT), which then subsequently lead to platelet activation and thrombus formation.20 As RTE sums up the coagulation aspects resulting in clotformation, the coagulation values differed significantly between the groups with physiological and pathological RTE as expected. Interestingly, male sex and trauma were significantly more frequent in patients with pathological RTE. This might be because of the acute coagulopathy following trauma,21 while the predisposition of male gender is most likely secondary because of the lower extremity injury pattern, which is more often found in young males. Fibrinogen has been demonstrated to be a major factor for microvascular thrombosis in a trauma setting.23 In addition, flap failure rates in patients with hyperfibrinogemia have been described as high as 20% by Kuo and coworkers. The same group on the other hand found that
Figure 2. Course of treatment with regards to RTE status and thrombotic complications. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Microsurgery DOI 10.1002/micr
258
Kolbenschlag et al.
Table 5. Results of Binary Logistic Regression for the Mentioned RTE Values Thrombotic event
Thrombotic flap loss
Influencing variables
P value
Odds ratio
P value
Odds ratio
AUC
ECPT MCF >65 pathological RTE FPR >43
0.61 0.82 0.51
1.3 1.1 1.4
0.13 0.036 0.003
4.97 3.75 7.90
0.60 0.63 0.66
In all models, multivariate analysis adjusted for Age, Sex, and Comorbidities was performed. ECPT, extrinsic coagulation pathway thrombus formation; MCF, mean clot firmness; FPR, functional fibrinogen to platelet ratio; RTE, rotational thromboelastometry; AUC, area under the curve.
this condition alone did not increase the rate of microvascular thrombosis in an animal model.24 Interestingly, in our series the fibrinogen levels were significantly higher in patients with flap loss, whereas the amount of thrombocytes did not differ significantly. This is backed up by the fact that the mean MCF value in ICF was elevated, also pointing towards the potential impact of functional fibrinogen. Hemoglobin was also found to be significantly lower in patients with pathological RTE in our study. Hill et al. found pre-operative anemia to be a predictor of thrombosis and free flap failure.22 They concluded that hemodilution might not prevent thrombotic complications, but instead even promote such events. In our study, preoperative hemoglobin levels were significantly lower in patients with pathological RTE findings. When looking at primary pedicle thrombosis, none of the included variables showed a significant influence. This might be because of the fact that most thromboembolic complications are secondary to technical errors or direct vessel trauma. Therefore, the fraction of thrombosis because of coagulatory disorders does not carry much weight and does not result in a significant difference. In patients without coagulatory disorders, the technical errors leading to thrombosis (e.g. kinking, technical failure of the anastomosis) can be corrected during revision surgery. In patients with such disorders, however, recurrent thrombosis cannot be prevented by surgical means and might reoccur despite such interventions, resulting in loss of the free flap. All but one of the thromboembolic flap losses occurred in patients with an ECPT-MCF of 66 and higher. We choose this cutoff within the physiological range as we hypothesized that the microvascular pedicle might be more prone to thrombosis because of its smaller diameter. Likely because of the sample size, this parameter did not prove to be a significant predictor of thrombotic flap loss (P 5 0.13). In a larger collective, an ECPT-MCF of 66 might function as a potential Microsurgery DOI 10.1002/micr
threshold. However, such a low cutoff value is likely to sacrifice specifity and lead to false positive results. Based on our findings, a hypercoagulable RTE is a better screening value for thrombotic flap loss (P 5 0.036) with a better predictive potential (AUC 5 0.63) and a nearly four times higher risk when compared to patients with physiological RTE (OR 5 3.75). In our calculation of the functional FPR, we found a significant higher rate of flap losses in the group with a ratio of >43 (P 5 0.003). This resembles the findings of Parker et al., where a ratio of 42 and higher was considered the threshold for thrombotic complications.19 Thus, a FPR of >43 seems to be an even stronger predictor of thrombotic flap loss (P 5 0.003) with a nearly eight times higher relative risk when compared to patients with physiological RTE (OR 5 7.9) and improved predictive potential (AUC 5 0.66). In addition, onset of thrombosis leading to flap loss was considerably later in patients with pathological RTE. In these patients, flap salvage was much less successful than in patients with physiological RTE. This is consistent with the literature, where Kroll and co-workers found that salvage attempts yielded muss worse results in late presentation of pedicle thrombosis.25 Wang et al. also reported similar findings with delayed presentation of thrombosis and a flap loss rate of 20% in patients with proven hypercoagulable state.3 This might be because of the fact that in case of technical errors such as exposed endothelium, the coagulatory cascade is immediately triggered, resulting in a platelet thrombus. On the other hand, in patients with systemic coagulatory disorders, some form of physiological suppression of coagulation might still be in order, being able to compensate the initial clot formation, e.g. by means of fibrinolysis. But, when these mechanisms are overwhelmed by a dysregulated coagulatory cascade, recurrent thromboembolic events take place, possibly resulting in delayed flap loss. Another possible explanation is that in many patients with pathological RTE, a continuous drip of unfractionated heparin was used. According to our internal standards, this treatment is discontinued after the sixth postoperative day. This change in the anticoagulatory regimen might also be partly responsible for those late flap losses. As demonstrated in a recent survey, 67% of all questioned surgeons administered some form of postoperative anticoagulation. However, both the applied agents and the duration they are given varies significantly between institutions.7 In our opinion, the subcutaneous application of lowmolecular heparin, as given for prevention of deep venous embolism anyway, is sufficient in most patients. This is especially true since intravenous application of
ROTEM in Microsurgery
unfractionated heparin has a much higher rate of bleeding complications when compared to subcutaneous application of low-molecular heparin.26,27 Therefore, identification of those patients that will benefit from an intensified anticoagulatory therapy is paramount. Based on our findings, patients with a pathological RTE (ECPT or ICPT MCF >72 mm or ICF >25 mm) or with a FPR of >43 are at a significantly higher risk for thrombotic flap loss (OR 3.75 and 7.9, respectively). In these patients, a PTT-guided anticoagulation with unfractionated heparin might be beneficial. Considering the delayed presentation of thromboembolic events, a prolonged duration of intensified anticoagulatory therapy should be considered. In addition, patients with pathological findings in RTE or with a known history of thrombosis, should undergo further investigation regarding the underlying pathology (e.g. protein C/S deficiency, factor V G1691A mutation Leiden, Antithrombin III deficiency or mutation, factor VIII elevation, prothrombin G20210A mutation, elevated homocysteine level, antiphospholipid syndrome, etc.) Several limitations apply to this study. Foremost, only a small number of patients developed pedicle thrombosis and in an even smaller group, thromboembolic flap loss occurred. Because of those small numbers, statistically significant predictors of thrombosis are hard to identify. Also, because of the retrospective nature of this work, potential confounders like the anticoagulation given might have some influence. To address these matters, a prospective multicenter study with different anticoagulatory regimen and RTE at different points in time would be necessary. CONCLUSION
In most patients, the anticoagulatory regimen is probably not the most crucial aspect of a successful free tissue transfer. In selected high-risks patients, however, it might have the potential to tip the scale. RTE and especially a FPR of >43 seem to be able to identify patients that are prone to thromboembolic complications following free tissue transfer and might be used as a point-ofcare tool to screen patients undergoing microvascular reconstruction. Especially in patients with recurrent thromboembolic events or even thromboembolic flap loss, RTE can verify pathological changes in coagulation, enabling a targeted therapy for those patients. Further studies are required to identify the best anticoagulatory regimen in correlation to RTE findings. ACKNOWLEDGMENTS
The authors thank Professor Hans Trampisch and Renate Klaaßen-Mielke of the Department of Medical Infor-
259
matics, Biometry and Epidemiology, University of Bochum for their help considering the statistical analysis.
REFERENCES 1. Wei F-C, Jain V, Celik N, Chen H-C, Chuang DC-C, Lin C-H. Have we found an ideal soft-tissue flap? An experience with 672 anterolateral thigh flaps. Plast Reconstr Surg 2002;109:2219–2226; discussion 2227–2230. 2. Bui DT, Cordeiro PG, Hu Q-Y, Disa JJ, Pusic A, Mehrara BJ. Free flap reexploration: Indications, treatment, and outcomes in 1193 free flaps. Plast Reconstr Surg 2007;119:2092–2100. 3. Wang TY, Serletti JM, Cuker A, McGrath J, Low DW, Kovach SJ, Wu LC. Free tissue transfer in the hypercoagulable patient: A review of 58 flaps. Plast Reconstr Surg 2012;129:443–453. 4. Yu P, Chang DW, Miller MJ, Reece G, Robb GL. Analysis of 49 cases of flap compromise in 1310 free flaps for head and neck reconstruction. Head Neck 2009;31:45–51. 5. Komatsu S, TamaiI S. Successful replantation of a completely cutoff thumb. Plast Reconstr Surg 1968. 6. Riva FM, Chen YC, Tan NC, Lin PY, Tsai YT, Chang HW, Kuo YR. The outcome of prostaglandin-E1 and dextran-40 compared to no antithrombotic therapy in head and neck free tissue transfer: analysis of 1,351 cases in a single center. Microsurgery 2012;32:339– 343. 7. Xipoleas G, Levine E, Silver L, Koch RM, Taub PJ. A survey of microvascular protocols for lower-extremity free tissue transfer I: perioperative anticoagulation. Ann Plast Surg 2007;59:311–315. 8. Srikanthan K, Viswanathan N, Yuen JC. Free-flap failure in thrombophilia: Case report and systematic review of the literature. Ann Plast Surg 2013 Feb 20 [Epub ahead of print]. 9. Warnecke IC, Kretschmer F, Bruner S, Frerichs O, Fansa H. [Hereditary thrombophilia in free microvascular flaps—A case report]. Handchir Mikrochir Plast Chir 2007;39:220–224. 10. McAnneny A, Durden F, Pearson GD, Tiwari P. Intra-flap thrombosis secondary to acute sickle crisis: a case report. Microsurgery 2012;32:585–587. 11. Hartert H. Blutgerinnungsstudien mit der Thromboelastographie, einem neuen Untersuchungsverfahren. Klin Wochenschr 1948;26: 577–583. 12. Bischof D, Dalbert S, Zollinger a, Ganter MT, Gantner MT, Hofer CK. Thrombelastography in the surgical patient. Minerva Anestesiol 2010;76:131–137. 13. Daudel F, Kessler U, Folly H, Lienert JS, Takala J, Jakob SM. Thromboelastometry for the assessment of coagulation abnormalities in early and established adult sepsis: a prospective cohort study. Critical care (London, England) 2009;13:R42. 14. Innerhofer P, Streif W, K€ uhbacher G, Fries D. Monitoring of perioperative dilutional coagulopathy using the ROTEM analyzer: Basic principles and clinical examples. Anasthesiol 2004;39:739–744. 15. Park MS, Martini WZ, Dubick Ma, Salinas J, Butenas S, Kheirabadi BS, Pusateri AE, Vos Ja, Guymon CH, Wolf SE, Mann KG, Holcomb JB. Thromboelastography as a better indicator of hypercoagulable state after injury than prothrombin time or activated partial thromboplastin time. J Trauma 2009;67:266–275; discussion 275–266. 16. Schaden E, Hoerburger D, Hacker S, Kraincuk P, Baron DM, Kozek-Langenecker S. Fibrinogen function after severe burn injury. Burns 2012;38:77–82. 17. Tauber H, Innerhofer P, Breitkopf R, Westermann I, Beer R, El Attal R, Strasak a, Mittermayr M. Prevalence and impact of abnormal ROTEM(R) assays in severe blunt trauma: results of the ‘Diagnosis and Treatment of Trauma-Induced Coagulopathy (DIA-TRETIC) study’. Br J Anaesth 2011;107:378–387. 18. McCrath DJ, Cerboni E, Frumento RJ, Hirsh AL, Bennett-Guerrero E. Thromboelastography maximum amplitude predicts postoperative thrombotic complications including myocardial infarction. Anesth Analg 2005;100:1576–1583.
Microsurgery DOI 10.1002/micr
260
Kolbenschlag et al.
19. Parker RJ, Eley KA, Von Kier S, Pearson O, Watt-Smith SR. Functional fibrinogen to platelet ratio using thromboelastography as a predictive parameter for thrombotic complications following free tissue transfer surgery: A preliminary study. Microsurgery 2012;32:512–519. 20. Furie B, Furie BC. Mechanisms of thrombus formation. N Engl J Med 2008;359:938–949. 21. Davenport R. Pathogenesis of acute traumatic coagulopathy. Transfusion 2013;53 Suppl 1:23S–27S. 22. Hill JB, Patel A, Del Corral Ga, Sexton KW, Ehrenfeld JM, Guillamondegui OD, Shack RB. Preoperative anemia predicts thrombosis and free flap failure in microvascular reconstruction. Ann Plast Surg 2012;69:364–367. 23. Khouri RK, Cooley BC, Kenna DM, Edstrom LE. Thrombosis of microvascular anastomoses in traumatized vessels: fibrin versus platelets. Plast Reconstr Surg 1990;86:110.
Microsurgery DOI 10.1002/micr
24. Kuo Y-R, Jeng S-F, Wu W-S, Lin C-J, Sacks JM, Yang KD. Hyperfibrinogenemia Alone Does Not Affect the Patency of Microvascular Anastomosis. Ann Plast Surg 2005;54:435–441. 25. Kroll SS, Schusterman MA, Reece GP, Miller MJ, Evans GR, Robb GL, Baldwin BJ. Timing of pedicle thrombosis and flap loss after free-tissue transfer. Plast Reconstr Surg 1996;98:1230–1233. 26. Costantino G, Ceriani E, Rusconi AM, Podda GM, Montano N, Duca P, Cattaneo M, Casazza G. Bleeding risk during treatment of acute thrombotic events with subcutaneous LMWH compared to intravenous unfractionated heparin: A systematic review. PLoS ONE 2012;7:e44553. 27. Erkens PM, Prins MH. Fixed dose subcutaneous low molecular weight heparins versus adjusted dose unfractionated heparin for venous thromboembolism. Cochrane database 2010: CD001100.
Thrombotic occlusion of the microvascular pedicle is the major reason for flap loss. Thus, identifying patients who are at risk for such events is paramount. Rotational thromboelastometry (RTE) is widely used to detect coagulopathy and hypercoagulable states. The aim of our study was to assess its diagnostic value in reconstructive microsurgery. In all 181 patients undergoing free tissue transfer at our department between February 2010 and November 2011 preoperative RTE was performed. In addition, coagulation values as well as patient’s demographic data, cause and localization of defect, type of flap and surgical revisions were recorded. The majority of patients was male (59.6%) with traumatic (59.7%) defects located on the lower extremity (60.3%). ALT was the most often used flap (35.9%). Preoperatively, 36.5% of patients had a hypercoagulable RTE (higher than physiological RTE values; intrinsic (ICPT) or extrinsic (ECPT) mean clot firmness (MCF) >72mm or functional fibrinogen (ICF) MCF >25mm). A total of 28 primary thrombosis of the microvascular pedicle occurred, 11 of those in-patients with a hypercoagulable state. Total flap loss rate because ofthrombosis was 7.7% (n 5 14). Both a hypercoagulable RTE assay and a functional fibrinogen to platelet ratio (FPR) of >43 (MCF value of ICF divided by the MCF value of ICPT) were significant predictors of thrombotic flap loss when performing multivariate binary logistic regression, co-factoring for age, sex, and comorbidities (p 5 0.036 and 0.003, respectively). RTE seems to be able to identify patients that are prone to thrombotic complicaC 2013 Wiley Periodicals, Inc. Microsurgery 34:253–260, 2014. tions and might be used as a screening tool. V
Free tissue transfer has had a significant impact on reconstructive surgery over the past decades. The techniques for microvascular tissue transfer have steadily evolved, resulting in high success rates.1 But even despite this tremendous progress, flap loss remains a looming threat with devastating consequences for both patients and surgeons. Flap loss most often occurs because of thrombosis of the microvascular pedicle and may be secondary to local and technical factors (e.g. intima damage, vessel kinking, technical errors regarding the anastomosis) or systemic influences (e.g. acquired or hereditary coagulation disorders like a posttraumatic hypercoagulable state or resistance to activated protein C).2–4 Therefore, the identification of patients with such risk factors and the modification of coagulation to prevent pedicle thrombosis has been an important topic since the early days of microsurgery.5 But even today, the perioperative management of anticoagulation in microvascular tissue transfer remains a highly controversial topic.6 Recent surveys show that a wide range of different anti1 BG University Clinic Bergmannsheil, Department of Plastic Surgery, Burn Center, Bochum, Germany 2 BG Trauma Center, Department of Plastic and Hand Surgery, Ludwigshafen, Germany 3 BG Trauma Center, Department of Anaesthesiology and Intensive Care Medicine, Ludwigshafen, Germany *Correspondence to: Jonas Kolbenschlag, M.D.; BG University Clinic Berg€rkle-de-la-Campmannsheil, Department of Plastic Surgery, Burn Center, Bu Platz 1, 44789 Bochum, Germany. Email: [email protected] Received 14 May 2013; Revision accepted 3 October 2013; Accepted 4 October 2013 Published online 20 October 2013 in Wiley Online Library (wileyonlinelibrary. com). DOI: 10.1002/micr.22199
Ó 2013 Wiley Periodicals, Inc.
coagulatory regimes is in place today.7 Such anticoagulatory therapy is seldom tailored to the specific patients risk profile, mostly because of the difficulties in identifying such patients preoperatively. While an immense amount of manpower, time, and technical utilities is dedicated to flap monitoring in the early postoperative period, less attention has been paid to the preoperative identification of potential high-risk patients. Although such close monitoring is essential for early recognition of pedicle thrombosis and potential flap salvage, a preoperative screening of the coagulatory state of these patients might allow for a targeted anticoagulatory therapy, aid in reducing such incidents and could therefore potentially improve the success rates of free tissue transfer even further. This is especially true as coagulatory disorders are quite common. Factor-V-Leiden Thrombophilia for example is prevalent in up to 15% of the European population.8,9 Other pro-thrombotic conditions such as sickle cell trait even occur in up to 1 in 3 persons in selected populations.10 As factor analysis and in-depth coagulation diagnostics are too costly to perform in every patient undergoing free tissue transfer, a global screening test for coagulatory disorders is needed. Rotational thromboelastometry (RTE) has first been described by Hartert in 1948.11 It generates a real-time image of in vitro clot formation and therefore takes into account not only the plasmatic components of coagulation but also its cellular elements and their interactions. By summing up these factors, it can act as a global screening test of the coagulatory state. RTE has seen an increasing use in various fields because of its point-of-care-
254
Kolbenschlag et al.
availability and real-time results. It is now widely used to detect coagulopathy and hypercoagulable states in perioperative and trauma care and has even been shown to predict myocardial infarction. Recently it was also used as a diagnostic tool in severe burns.12–18 However, very little data exist about the diagnostic value of RTE in microvascular tissue transfer. To our knowledge, only a small preliminary study dealing with this subjected has been published up to now. In 29 patients undergoing free tissue transfer for head and neck reconstruction, Parker et al. found that certain RTE values might be able to predict thrombotic complications in microvascular surgery.19 Therefore, the aim of our study was to assess the diagnostic value of RTE regarding thrombotic complications in reconstructive microsurgery.
ter. Maximal clot firmness (MCF) was defined as the maximum diameter of the clot, measured in millimeters. A hypercoagulabe state leads to increased activation of coagulation, resulting in a larger thrombus and therefore an increased MCF. The physiological range of MCF for ECPT and ICPT is 50–72 and 9–25 mm for ICF as supplied by the manufacturer. The physiological ranges for the other values are also given in Figure 1. In this graphic display of results, time was plotted against mm of clot thickness, resulting in an image of the actual thrombus and its behavior over time. All three activated tests for extrinsic and intrinsic coagulation pathway of thrombus formation and the isolated contribution of functional fibrinogen to the clot are displayed below each other in Figure 1, with the graphic display on the left and numeric results on the right side.
PATIENTS AND METHODS
Hypercoagulable RTE and Functional Fibrinogen to Platelet Ratio
A retrospective review of all patients undergoing free tissue transfer at our department between February 2010 and November 2011 was conducted. In all of the 181 consecutive patients included, pre-operative RTE was R -Device (Tem Internaperformed using the ROTEMV tional GmbH, Munich, Germany) as a standard of care. Rotational Thromboelastometry
RTE is based on an in vitro clotting process. It generates a real-time image of in vitro clot formation and therefore takes into account not only the plasmatic components of coagulation but also its cellular elements and their interactions. By summing up these factors, it can act as a global screening test of the coagulatory state. For RTE-tests, blood samples were given into a cup, which housed a rotating pin. Under continuous circulation of the sample, the blood began to clot. Although it is possible to measure native thrombus formation without any additional activation, activated tests offer improved reliability and reproducibility. For such tests, by activation of either the extrinsic coagulation pathway thrombus formation (ECPT) or intrinsic coagulation pathway thrombus formation (ICPT), thrombus formation was initiated by adding tissue factor for the extrinsic and partial thromboplastin for the intrinsic pathway. In addition, by stimulating the extrinsic pathway and adding a platelet-inhibiting reagent (Cytochalacin), the isolated contribution of functional fibrinogen to the clot was detected (ICF; isolated contribution of fibrinogen). After activation, the in vitro clot formation began and was continuously measured for all three tests to generate a real-time image of several clotting attributes. Coagulation time (CT) was defined as the duration between the initiation of clot formation and the first visible clotting. The time after which the clot achieved a diameter of 20 mm was referred to as clot-formation-time (CFT). The a-angle characterized the kinetics with which the clot gained firmness and diameMicrosurgery DOI 10.1002/micr
In the study presented here, a complete RTE, as performed for each patient, consisted of those three activated tests. On the preoperative day, blood was drawn for standard laboratory work-up, including a coagulation screening and RTE measurements. RTE was considered hypercoagulable when the MCF for the ECPT and ICPT was higher than 72 mm, or higher than 25 mm for ICF. If the RTE values were within the physiological range as supplied by the manufacturer (Fig. 1), RTE was considered physiological. For calculating the functional fibrinogen to platelet ratio (FPR) the MCF of ICF was divided by the MCF of ICPT, resulting in a numeric ratio. For the last analyzed value we chose to use a cutoff value within the physiological range (66 mm) for ECPT-MCF (Mean Firmness of the clot in the extrinsic pathway test). In addition preoperative fibrinogen-, thrombocyte-, hemoglobin-levels as well as PTT and AntiThrombin III activity (AT III) were recorded. We also recorded patient’s sex, age, comorbidities, cause and localization of defect, type of flap used and surgical revisions with an emphasis on thromboembolic events. No pre-operative anticoagulation was given other than the previous medication, postoperative anticoagulation treatment was administered at discretion of the surgeon, either as continuous infusion of unfractionated heparin (15.000U/24h) or as low-molecular heparin subR 0.4 ml, two times a day). cutaneously (ClexaneV Statistical analysis of the data was performed utilizing SPPS 21 for Mac (IBM Corporation) and Excel 2011 for Mac (Microsoft Corporation). For comparison between groups, student’s t-test was used, whereas the correlation of influencing variables with thromboembolic complications and flap loss was calculated using binary logistic regression. Additionally, the odds ratio was calculated for the mentioned risk factors.
ROTEM in Microsurgery
255
R ), ICPT (INTEMV R ), and ICF Figure 1. Hypercoagulable RTE in a patient with thrombotic flap loss. From top to bottom, ECPT (EXTEMV R ) were depicted. The graph represents the actual in vivo clot formation. The clot’s diameter (isolated contribution of fibrinogen; FIBTEMV in mm (MCF) was plotted against time in minutes. Note the elevated MCF in all three tests. Also, the physiological range for each value is given in brackets (CT, clotting time; CFT, clot formation time; ML, maximum lysis; A10 / A20, clot diameter at 10 / 20 minutes). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
The area under the curve (AUC) was calculated to display the predictive potential of certain RTE values. Results are displayed as mean 6 SD. A significance level of P < 0.05 was assumed.
RESULTS
A total of 181 patients with free flaps were included, of which 108 (59.6%) were male. Mean age at the time of operation was 50.2 614.1 years. Thirty-seven per cent of patients were active smokers, 35% had a history of hypertension, 17% were obese (BMI > 25), 9% suffered from peripheral arterial disease (PAD), 8% from diabetes. A history of thromboembolic events was present in 4% of all patients. Soft-tissue defects necessitating free-flap coverage were most often because of trauma (59.7%) and malignancies (24.9%). Most defects were located on the lower extremity (60.3%) followed by the trunk (21.5%), upper extremity (16.6%), and head/neck (1.6%; Table 1). The most often used type of flap was the antero-lateral thigh flap (ALT) in 35.9%, followed by the
Table 1. Distribution of Comorbidities, Cause of Defect and Location of Defect Among the Included Patients Comorbidity/Cause of Defect/Localization Smoking Hypertension Obesity (BMI >25) Diabetes History of thrombosis Peripheral Arterial Disease (PAD) Trauma Malignancy Infection Burns Chronic Ulcer Lower extremity Trunk Upper extremity Head/Neck
No. of patients
%
67 64 31 14 7 16
37% 35.3% 17% 7.7% 3.8% 8.8%
108 45 12 9 7 109 39 30 3
59.7% 24.9% 6.6% 4.9% 3.9% 60.3% 21.5% 16.6% 1.6%
latissimus dorsi flap (24.9%), deep inferior epigastric artery perforator flap (DIEP; 16.5%), parascapular flap (11.6), free gracilis flap (4.5%), fibula flap (3.9%), lateral Microsurgery DOI 10.1002/micr
256
Kolbenschlag et al.
arm flap (1.6%), and superficial gluteal artery perforator flap (S-GAP; 1.1%; Table 2). Mean duration of hospital stay was 46.8 638.69 days (Median 34 days, Range 9–252 days). In 66 patients (36.5%) the pre-operative RTE was deemed hypercoagulable (ECPT or ICPT MCF >72 mm or ICF MCF >25 mm), whereas in the remaining patients the values were within standard range. In all but one patient, if one value (ECPT, ICPT, ICF) was within pathological range the others were too. In this patient, ECPT and ICF were elevated while ICPT was borderline normal with 72 mm. A hypocoagulable state was not seen. The age distribution between patients with pathological and physiological RTE did not significantly differ. Patients with a hypercoagulable RTE were much more likely to be men (P 5 0.0003) with traumatic defects (P 5 0.0076) of the lower extremity (P 5 0.0015). Hypertension, obesity, and history of smoking were not significantly different between the two groups (P 5 0.6, 0.22 and1, respectively). Considering the laboratory measurements, concentrations of fibrinogen and thrombocytes were significantly higher in patients with hypercoagulable RTE (P 5 0.0001 for both). AT III activity was also significantly higher in
Table 2. Distribution of Free Flaps Among the Included Patients Type of flap
No. of patients
%
65 45 30 21 8 7 3 2
35.9% 24.9% 16.5% 11.6% 4.5% 3.9% 1.6% 1.1%
ALT Latissimus dorsi DIEP Parascapular Gracilis Fibula Lateral Arm S-GAP
ALT, anterior lateral thigh flap; DIEP, deep inferior epigastric artery perforator flap; S-GAP, superior gluteal artery perforator flap.
these patients (P 5 0.001), whereas the partial thromboplastin time was longer (P 5 0.019). Hemoglobin concentration on the other hand was higher in patients with a physiological RTE (Table 3). Overall, 28 pedicle thrombosis occurred during the postoperative course of treatment (15.5%). Venous thromboembolisms were found in 15 patients, whereas arterial thrombosis was found in six patients. In seven patients, both arterial and venous vessels were occluded by thromboembolisms. Seventeen of these thrombosis occurred in the 115 patients with physiological RTE (15%), whereas the remaining eleven were found in the 66 patients with a hypercoagulable RTE (16.6%). There was no significant difference in the onset of primary thrombosis between the subgroups (P 5 0.19). The total rate of free flap loss because of thrombotic events was 7.7%, with 14 free flaps being lost to thrombosis. Nine of these flap losses occurred in patients with a pathological RTE (13.6%), whereas the other five were seen in the remaining 115 patients (4.3%). In the group with thrombotic flap losses, fibrinogen levels were significantly higher than in patients without flap losses (P 5 0.024), while there was no such difference regarding thrombocyte level (P 5 0.85; Table 4). In addition, the first thrombosis in flaps which were subsequently lost to thrombotic events occurred a mean of 5.6 days later in patients with hypercoagulable RTE when compared to those with physiological ones (3.6 63.2 vs. 9.2 68 days). Also, the salvage rate in flaps affected by thrombotic complications was considerably lower in the group with pathological RTE findings. While 12 out of the 17 flaps (71%) with thromboembolic complications could be salvaged in patients with physiological RTE, this was only possible in two out of 11 (18%) flaps in the hypercoagulable group (Fig. 2).
Table 3. Comparison of Patients with Pathological and Physiological RTE
Age Sex (male / female) Trauma (yes / no) Malignancy (yes / no) Hypertension (yes / no) Smoking history(yes / no) Lower extremity(yes / no) Obesity (yes / no) Fibrinogen (mg/dl) Thrombocytes (Tsd./ul) PTT (s.) Hemoglobin (g/dl) Antithrombin III (%) PTT, partial thromboplastin time.
Microsurgery DOI 10.1002/micr
Physiological RTE (n 5 115)
Hypercoagulable RTE (n 5 66)
P-value
51.07 613,4 57/58 60/55 40/75 41/74 43/72 59/56 23/92 320.8 610.6 292.35 6104.3 34.1 66.9 12.6 61.8 78.77 614
48.82 615,28 51/15 48/18 6/60 21/45 24/42 50/16 8/58 461.9 694.9 469.52 6175.6 36.47 65.6 10.9 61.7 87.7616.5
0.3 0.0003 0.0076 0.0001 0.6 1 0.0015 0.22 0.0001 0.0001 0.019 0.0001 0.001
ROTEM in Microsurgery
Thirteen out of the 14 thromboembolic flap losses (92%) occurred in the 134 patients with an ECPT-MCF of 66 and higher (9.7% flap loss rate). When calculating the functional FPR using the RTE device (ICF/ICPT ratio, FPR), we found that only 38 patients had a ratio of >43 (21% of all patients). Out of the 14 flap losses, eight (53%) occurred in these patients. All three values (pathological RTE, ECPT-MCF>65, FPR >43) were introduced into multivariate binary logistic regression models adjusting for sex, age, and comorbidities (PAD, history of smoking or thrombosis, diabetes, arterial hypertension, obesity). Considering primary thrombotic events, none of the variables was of significant influence. However, both a pathological RTE and a FPR >43 were strong predictors of thrombotic flap loss (P 5 0.036 and 0.003, respectively; Table 5). None of the other computed values showed a significant correlation with thrombotic flap losses. Table 4. Comparison of Coagulation Values in Patients with and Without Thrombotic Flap Loss
ECPT (50–72mm) ICPT (50–72mm) ICF (9–25mm) Fibrinogen (mg/dl) Thrombocytes (Tsd./ul)
No flaploss (n 5 167)
Flaploss (n 5 14)
P-value
69,9 6 6,14 67,62 6 6,93 23,46 6 10,59 362 6 120 357 6 160.7
73,43 6 7,08 70,93 6 5,42 26,93 6 9,78 441 6 112 349.1 6 138.1
0,074 0,08 0,23 0,024 0.85
ECPT, extrinsic coagulation pathway thrombus formation; ICPT, intrinsic coagulation pathway thrombus formation; ICF, isolated contribution of fibrinogen.
257
DISCUSSION
RTE was first described over 60 years ago, but had not found broad clinical application in the beginning.11 This was mostly because of the complex and interference-prone nature of these early devices. In recent times, RTE devices have become much more user friendly because of automation of the performed test and improved data analysis and handling capabilities. The R device (Tem International GmbH, Munich, ROTEMV Germany) is a point-of-care device capable of performing four parallel measurements, resulting in easy and fast to obtain coagulation measurements. ECPT and ICPT depict activation of the coagulatory cascade either triggered by tissue factor (ECPT) or partial thromboplastin (ICPT), which then subsequently lead to platelet activation and thrombus formation.20 As RTE sums up the coagulation aspects resulting in clotformation, the coagulation values differed significantly between the groups with physiological and pathological RTE as expected. Interestingly, male sex and trauma were significantly more frequent in patients with pathological RTE. This might be because of the acute coagulopathy following trauma,21 while the predisposition of male gender is most likely secondary because of the lower extremity injury pattern, which is more often found in young males. Fibrinogen has been demonstrated to be a major factor for microvascular thrombosis in a trauma setting.23 In addition, flap failure rates in patients with hyperfibrinogemia have been described as high as 20% by Kuo and coworkers. The same group on the other hand found that
Figure 2. Course of treatment with regards to RTE status and thrombotic complications. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Microsurgery DOI 10.1002/micr
258
Kolbenschlag et al.
Table 5. Results of Binary Logistic Regression for the Mentioned RTE Values Thrombotic event
Thrombotic flap loss
Influencing variables
P value
Odds ratio
P value
Odds ratio
AUC
ECPT MCF >65 pathological RTE FPR >43
0.61 0.82 0.51
1.3 1.1 1.4
0.13 0.036 0.003
4.97 3.75 7.90
0.60 0.63 0.66
In all models, multivariate analysis adjusted for Age, Sex, and Comorbidities was performed. ECPT, extrinsic coagulation pathway thrombus formation; MCF, mean clot firmness; FPR, functional fibrinogen to platelet ratio; RTE, rotational thromboelastometry; AUC, area under the curve.
this condition alone did not increase the rate of microvascular thrombosis in an animal model.24 Interestingly, in our series the fibrinogen levels were significantly higher in patients with flap loss, whereas the amount of thrombocytes did not differ significantly. This is backed up by the fact that the mean MCF value in ICF was elevated, also pointing towards the potential impact of functional fibrinogen. Hemoglobin was also found to be significantly lower in patients with pathological RTE in our study. Hill et al. found pre-operative anemia to be a predictor of thrombosis and free flap failure.22 They concluded that hemodilution might not prevent thrombotic complications, but instead even promote such events. In our study, preoperative hemoglobin levels were significantly lower in patients with pathological RTE findings. When looking at primary pedicle thrombosis, none of the included variables showed a significant influence. This might be because of the fact that most thromboembolic complications are secondary to technical errors or direct vessel trauma. Therefore, the fraction of thrombosis because of coagulatory disorders does not carry much weight and does not result in a significant difference. In patients without coagulatory disorders, the technical errors leading to thrombosis (e.g. kinking, technical failure of the anastomosis) can be corrected during revision surgery. In patients with such disorders, however, recurrent thrombosis cannot be prevented by surgical means and might reoccur despite such interventions, resulting in loss of the free flap. All but one of the thromboembolic flap losses occurred in patients with an ECPT-MCF of 66 and higher. We choose this cutoff within the physiological range as we hypothesized that the microvascular pedicle might be more prone to thrombosis because of its smaller diameter. Likely because of the sample size, this parameter did not prove to be a significant predictor of thrombotic flap loss (P 5 0.13). In a larger collective, an ECPT-MCF of 66 might function as a potential Microsurgery DOI 10.1002/micr
threshold. However, such a low cutoff value is likely to sacrifice specifity and lead to false positive results. Based on our findings, a hypercoagulable RTE is a better screening value for thrombotic flap loss (P 5 0.036) with a better predictive potential (AUC 5 0.63) and a nearly four times higher risk when compared to patients with physiological RTE (OR 5 3.75). In our calculation of the functional FPR, we found a significant higher rate of flap losses in the group with a ratio of >43 (P 5 0.003). This resembles the findings of Parker et al., where a ratio of 42 and higher was considered the threshold for thrombotic complications.19 Thus, a FPR of >43 seems to be an even stronger predictor of thrombotic flap loss (P 5 0.003) with a nearly eight times higher relative risk when compared to patients with physiological RTE (OR 5 7.9) and improved predictive potential (AUC 5 0.66). In addition, onset of thrombosis leading to flap loss was considerably later in patients with pathological RTE. In these patients, flap salvage was much less successful than in patients with physiological RTE. This is consistent with the literature, where Kroll and co-workers found that salvage attempts yielded muss worse results in late presentation of pedicle thrombosis.25 Wang et al. also reported similar findings with delayed presentation of thrombosis and a flap loss rate of 20% in patients with proven hypercoagulable state.3 This might be because of the fact that in case of technical errors such as exposed endothelium, the coagulatory cascade is immediately triggered, resulting in a platelet thrombus. On the other hand, in patients with systemic coagulatory disorders, some form of physiological suppression of coagulation might still be in order, being able to compensate the initial clot formation, e.g. by means of fibrinolysis. But, when these mechanisms are overwhelmed by a dysregulated coagulatory cascade, recurrent thromboembolic events take place, possibly resulting in delayed flap loss. Another possible explanation is that in many patients with pathological RTE, a continuous drip of unfractionated heparin was used. According to our internal standards, this treatment is discontinued after the sixth postoperative day. This change in the anticoagulatory regimen might also be partly responsible for those late flap losses. As demonstrated in a recent survey, 67% of all questioned surgeons administered some form of postoperative anticoagulation. However, both the applied agents and the duration they are given varies significantly between institutions.7 In our opinion, the subcutaneous application of lowmolecular heparin, as given for prevention of deep venous embolism anyway, is sufficient in most patients. This is especially true since intravenous application of
ROTEM in Microsurgery
unfractionated heparin has a much higher rate of bleeding complications when compared to subcutaneous application of low-molecular heparin.26,27 Therefore, identification of those patients that will benefit from an intensified anticoagulatory therapy is paramount. Based on our findings, patients with a pathological RTE (ECPT or ICPT MCF >72 mm or ICF >25 mm) or with a FPR of >43 are at a significantly higher risk for thrombotic flap loss (OR 3.75 and 7.9, respectively). In these patients, a PTT-guided anticoagulation with unfractionated heparin might be beneficial. Considering the delayed presentation of thromboembolic events, a prolonged duration of intensified anticoagulatory therapy should be considered. In addition, patients with pathological findings in RTE or with a known history of thrombosis, should undergo further investigation regarding the underlying pathology (e.g. protein C/S deficiency, factor V G1691A mutation Leiden, Antithrombin III deficiency or mutation, factor VIII elevation, prothrombin G20210A mutation, elevated homocysteine level, antiphospholipid syndrome, etc.) Several limitations apply to this study. Foremost, only a small number of patients developed pedicle thrombosis and in an even smaller group, thromboembolic flap loss occurred. Because of those small numbers, statistically significant predictors of thrombosis are hard to identify. Also, because of the retrospective nature of this work, potential confounders like the anticoagulation given might have some influence. To address these matters, a prospective multicenter study with different anticoagulatory regimen and RTE at different points in time would be necessary. CONCLUSION
In most patients, the anticoagulatory regimen is probably not the most crucial aspect of a successful free tissue transfer. In selected high-risks patients, however, it might have the potential to tip the scale. RTE and especially a FPR of >43 seem to be able to identify patients that are prone to thromboembolic complications following free tissue transfer and might be used as a point-ofcare tool to screen patients undergoing microvascular reconstruction. Especially in patients with recurrent thromboembolic events or even thromboembolic flap loss, RTE can verify pathological changes in coagulation, enabling a targeted therapy for those patients. Further studies are required to identify the best anticoagulatory regimen in correlation to RTE findings. ACKNOWLEDGMENTS
The authors thank Professor Hans Trampisch and Renate Klaaßen-Mielke of the Department of Medical Infor-
259
matics, Biometry and Epidemiology, University of Bochum for their help considering the statistical analysis.
REFERENCES 1. Wei F-C, Jain V, Celik N, Chen H-C, Chuang DC-C, Lin C-H. Have we found an ideal soft-tissue flap? An experience with 672 anterolateral thigh flaps. Plast Reconstr Surg 2002;109:2219–2226; discussion 2227–2230. 2. Bui DT, Cordeiro PG, Hu Q-Y, Disa JJ, Pusic A, Mehrara BJ. Free flap reexploration: Indications, treatment, and outcomes in 1193 free flaps. Plast Reconstr Surg 2007;119:2092–2100. 3. Wang TY, Serletti JM, Cuker A, McGrath J, Low DW, Kovach SJ, Wu LC. Free tissue transfer in the hypercoagulable patient: A review of 58 flaps. Plast Reconstr Surg 2012;129:443–453. 4. Yu P, Chang DW, Miller MJ, Reece G, Robb GL. Analysis of 49 cases of flap compromise in 1310 free flaps for head and neck reconstruction. Head Neck 2009;31:45–51. 5. Komatsu S, TamaiI S. Successful replantation of a completely cutoff thumb. Plast Reconstr Surg 1968. 6. Riva FM, Chen YC, Tan NC, Lin PY, Tsai YT, Chang HW, Kuo YR. The outcome of prostaglandin-E1 and dextran-40 compared to no antithrombotic therapy in head and neck free tissue transfer: analysis of 1,351 cases in a single center. Microsurgery 2012;32:339– 343. 7. Xipoleas G, Levine E, Silver L, Koch RM, Taub PJ. A survey of microvascular protocols for lower-extremity free tissue transfer I: perioperative anticoagulation. Ann Plast Surg 2007;59:311–315. 8. Srikanthan K, Viswanathan N, Yuen JC. Free-flap failure in thrombophilia: Case report and systematic review of the literature. Ann Plast Surg 2013 Feb 20 [Epub ahead of print]. 9. Warnecke IC, Kretschmer F, Bruner S, Frerichs O, Fansa H. [Hereditary thrombophilia in free microvascular flaps—A case report]. Handchir Mikrochir Plast Chir 2007;39:220–224. 10. McAnneny A, Durden F, Pearson GD, Tiwari P. Intra-flap thrombosis secondary to acute sickle crisis: a case report. Microsurgery 2012;32:585–587. 11. Hartert H. Blutgerinnungsstudien mit der Thromboelastographie, einem neuen Untersuchungsverfahren. Klin Wochenschr 1948;26: 577–583. 12. Bischof D, Dalbert S, Zollinger a, Ganter MT, Gantner MT, Hofer CK. Thrombelastography in the surgical patient. Minerva Anestesiol 2010;76:131–137. 13. Daudel F, Kessler U, Folly H, Lienert JS, Takala J, Jakob SM. Thromboelastometry for the assessment of coagulation abnormalities in early and established adult sepsis: a prospective cohort study. Critical care (London, England) 2009;13:R42. 14. Innerhofer P, Streif W, K€ uhbacher G, Fries D. Monitoring of perioperative dilutional coagulopathy using the ROTEM analyzer: Basic principles and clinical examples. Anasthesiol 2004;39:739–744. 15. Park MS, Martini WZ, Dubick Ma, Salinas J, Butenas S, Kheirabadi BS, Pusateri AE, Vos Ja, Guymon CH, Wolf SE, Mann KG, Holcomb JB. Thromboelastography as a better indicator of hypercoagulable state after injury than prothrombin time or activated partial thromboplastin time. J Trauma 2009;67:266–275; discussion 275–266. 16. Schaden E, Hoerburger D, Hacker S, Kraincuk P, Baron DM, Kozek-Langenecker S. Fibrinogen function after severe burn injury. Burns 2012;38:77–82. 17. Tauber H, Innerhofer P, Breitkopf R, Westermann I, Beer R, El Attal R, Strasak a, Mittermayr M. Prevalence and impact of abnormal ROTEM(R) assays in severe blunt trauma: results of the ‘Diagnosis and Treatment of Trauma-Induced Coagulopathy (DIA-TRETIC) study’. Br J Anaesth 2011;107:378–387. 18. McCrath DJ, Cerboni E, Frumento RJ, Hirsh AL, Bennett-Guerrero E. Thromboelastography maximum amplitude predicts postoperative thrombotic complications including myocardial infarction. Anesth Analg 2005;100:1576–1583.
Microsurgery DOI 10.1002/micr
260
Kolbenschlag et al.
19. Parker RJ, Eley KA, Von Kier S, Pearson O, Watt-Smith SR. Functional fibrinogen to platelet ratio using thromboelastography as a predictive parameter for thrombotic complications following free tissue transfer surgery: A preliminary study. Microsurgery 2012;32:512–519. 20. Furie B, Furie BC. Mechanisms of thrombus formation. N Engl J Med 2008;359:938–949. 21. Davenport R. Pathogenesis of acute traumatic coagulopathy. Transfusion 2013;53 Suppl 1:23S–27S. 22. Hill JB, Patel A, Del Corral Ga, Sexton KW, Ehrenfeld JM, Guillamondegui OD, Shack RB. Preoperative anemia predicts thrombosis and free flap failure in microvascular reconstruction. Ann Plast Surg 2012;69:364–367. 23. Khouri RK, Cooley BC, Kenna DM, Edstrom LE. Thrombosis of microvascular anastomoses in traumatized vessels: fibrin versus platelets. Plast Reconstr Surg 1990;86:110.
Microsurgery DOI 10.1002/micr
24. Kuo Y-R, Jeng S-F, Wu W-S, Lin C-J, Sacks JM, Yang KD. Hyperfibrinogenemia Alone Does Not Affect the Patency of Microvascular Anastomosis. Ann Plast Surg 2005;54:435–441. 25. Kroll SS, Schusterman MA, Reece GP, Miller MJ, Evans GR, Robb GL, Baldwin BJ. Timing of pedicle thrombosis and flap loss after free-tissue transfer. Plast Reconstr Surg 1996;98:1230–1233. 26. Costantino G, Ceriani E, Rusconi AM, Podda GM, Montano N, Duca P, Cattaneo M, Casazza G. Bleeding risk during treatment of acute thrombotic events with subcutaneous LMWH compared to intravenous unfractionated heparin: A systematic review. PLoS ONE 2012;7:e44553. 27. Erkens PM, Prins MH. Fixed dose subcutaneous low molecular weight heparins versus adjusted dose unfractionated heparin for venous thromboembolism. Cochrane database 2010: CD001100.
