FREE FLAP MONITORING USING AN IMPLANTABLE ANASTOMOTIC VENOUS FLOW COUPLER: ANALYSIS OF 119 CONSECUTIVE ABDOMINAL-BASED FREE FLAPS FOR BREAST RECONSTRUCTION STEVE J. KEMPTON, M.D., SAMUEL O. POORE, M.D., Ph.D., JENNY T. CHEN, M.D., and AHMED M. AFIFI, M.D.*
Purpose: The purpose of this study is to evaluate the use of the venous anastomotic Flow Coupler in monitoring free flaps used for breast reconstruction in a consecutive series of patients. Methods: Retrospective data were collected on patients undergoing free flap breast reconstruction from May 2012 to March 2014. The venous anastomotic Flow Coupler was used in the first 85 flaps and a non-flow Coupler with clinical and external Doppler monitoring alone in the subsequent 34 flaps. Data collected included patient age, BMI, prior radiation, flap type, intra- and postoperative Flow Coupler events, along with rates of flap take back, salvage, and failure. Proportion data were compiled and statistically analyzed. Results: One hundred nineteen consecutive abdominal based breast reconstruction free flaps were performed. The overall flap failure rate was 4.2% (4.7% Flow Coupler; 2.9% in non-flow Coupler; P 5 1.0). The Flow Coupler demonstrated 100% sensitivity in the intra- and postoperative settings. A positive predictive value of 36% was noted intraoperatively which was significantly higher compared to the non-flow Coupler group (P 5 0.015). Vessel thrombosis occurred in 17.6% of Flow Coupler flaps, which was significantly higher when compared to the non-flow Coupler (2.9%; P 5 0.038). Conclusions: The Flow Coupler is a sensitive method to confirm patency of a microsurgical anastomosis. However, there is a high false positive rate in both the intraoperative and postoperative settings resulting in frequent intraoperative maneuvers to amend the cause of signal loss. Additionally, the Flow Coupler resulted C 2014 Wiley Periodicals, Inc. Microsurgery in significantly more vascular thrombotic events when compared to the non-flow Coupler. V 35:337–344, 2015.
There
is substantial variation in flap monitoring techniques following free tissue transfer including surface temperature recording, transcutaneous laser and pencil Doppler, PO2 monitoring, visible white light spectroscopy, intravenous fluorescein, photoplethysmography, and implantable Doppler monitoring.1–7 Clinical observation remains the gold standard assessment of flap viability, but is labor intensive and requires skill and experience.7,8 The ideal device should be safe for the patient and the flap, allow for continuous monitoring, be sensitive and reliable in diagnosing both venous and arterial insufficiency, and be applicable to all flap types.9 In the 1980s, The Cook–Swartz implantable Doppler system was the first to enable direct anastomotic Doppler monitoring, consisting of a Doppler probe attached near the venous anastomosis using a silicone leash.10 This method of flap monitoring has demonstrated reliability is multiple series, however, requires additional operative steps and can be difficult to apply.1,3,11–13 The advent of the venous anastomotic Coupler by Synovis Life Technologies, Inc. has changed the way the venous anastomosis is performed and Division of Plastic and Reconstructive Surgery, University of Wisconsin Hospital and Clinics, Madison, WI Grant sponsor: Division of Plastic and Reconstructive Surgery at University of Wisconsin Hospital and Clinics. *Correspondence to: Ahmed M. Afifi, M.D., Division of Plastic and Reconstructive Surgery, Department of Surgery, University of Wisconsin Hospital and Clinics, 600 Highland Avenue, G5/361 Clinical Science Center, Mail code 3236, Madison, WI 53792. E-mail: [email protected] Received 3 June 2014; Revision accepted 26 September 2014; Accepted 3 October 2014 Published online 21 October 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/micr.22341 Ó 2014 Wiley Periodicals, Inc.
has been shown to be both safe and effective when compared to hand sewn techniques.14 In 2010, the Flow Coupler was introduced by Synovis Life Technologies, Inc., presenting a fusion of the venous anastomotic Coupler and a 20 MHz ultrasonic microdoppler probe. While the Flow Coupler is being used in clinic microsurgical practice, published date on the use of this device is limited.15 The aim of this work is to evaluate the use of the Flow Coupler in monitoring free flaps used for breast reconstruction in a consecutive series of patients, and describe intra- and postoperative events. PATIENTS AND METHODS
In December of 2011, the two senior authors (AA, SP) initiated the use of the venous anastomotic Flow Coupler to serve as an adjunct to clinical and external Doppler free flap monitoring (Fig. 1) (Synovis Micro Companies Alliance Inc., Birmingham, AL). From May of 2013, a Synovis Life Technologies, Inc. Coupler without an implantable Doppler probe (non-flow Coupler) was used in all subsequent abdominal-based free flaps. In these patients, external (pencil) Doppler examination and clinical assessment alone was used for free flap monitoring. Data included the occurrence, cause, and management of all intraoperative or postoperative events such as signal loss or flap vascular compromise. Following Institutional Review Board approval, hospital records, operative reports, and prospectively collected perioperative flap monitoring events were reviewed. Demographic data collected included patient age, BMI, and prior radiation. Perioperative data included type of
338
Kempton et al.
Figure 1. Intraoperative photo of venous anastomotic flow coupler with wire passing under the artery to exit the chest wall laterally.
abdominal-based flap (muscle sparing transverse rectus abdominis myocutaneous [MS-TRAM], deep inferior epigastric perforator [DIEP], superifical inferior epigastric artery [SIEA]), recipient vessel used for anastomosis, intraoperative and postoperative Flow Coupler events, along with rate of flap take back, salvage, and failure. After the first event of venous kinking that was found to be caused by traction from the Flow Coupler wire, the location of the internal mammary vein (IMV) in relation to the internal mammary artery (IMA), the course of the Flow Coupler wire, and wire exit point (medial vs. lateral) on the chest wall was additionally recorded. Application of Venous Anastomotic Flow Coupler
Abdominal-based free flap harvest and recipient vessel preparation was done simultaneously. The preferred recipient vessels were the IMA and IMV. Venous anastomosis always preceded arterial anastomosis. The recipient and donor veins were clamped separately. A vesselmeasuring gauge was used to determine Flow Coupler size. Donor and recipient veins were anastomosed using the Flow Coupler in the usual fashion. The arterial anastomosis was then completed using interrupted 8-0 nylon sutures. The Doppler wire was passed out percutaneously through a stab incision close to the IMF medially or laterally. In the setting of bilateral reconstruction, the above was repeated on the contralateral side.
operatively, the flap vessels and the anastomosis were examined, often requiring removal of the flap insetting sutures (between the flap and the chest wall). If there was good flow in the vessels, the flap was repositioned and the sutures used for inset were replaced. Thrombotic and occlusion problems were dealt with accordingly. The Doppler signals were continuously monitored following anastomosis until the patient left the operating room. Patients were transferred to the regular unit where flaps were monitored every hour for the first 24 hours, every 2 hours for the next 48 hours, then every 4 hours until discharge. Flap monitoring consisted of clinical examination (capillary refill, temperature, blanching, and skin turgor), Flow Coupler signal monitoring if the Flow Coupler was used, and external Doppler examination. Operative take back was defined as return to the operating room for flap vascular compromise and therefore did not include evacuation of hematoma in the setting of a viable flap, debridement of non-viable mastectomy skin flaps, or any other secondary procedures. Data Analysis
Flap monitoring during the intraoperative and postoperative periods were analyzed separately. Only one signal loss in the intraoperative and one signal loss in the postoperative setting were used for the analysis. A true positive event was defined as loss of signal and a compromised flap. A false positive event was the loss of a signal in the setting of a viable flap. A true negative event was defined as no loss of signal and a viable flap. A false negative event was no loss of signal and a compromised flap. Positive and negative predictive values (PPV/NPV), sensitivity, and specificity were calculated to describe the performance of the Flow Coupler. Student’s t-test and Fischer exact test were used to compare data between breast reconstruction cases with use of the Flow Coupler and cases that used the nonflow Coupler with clinical monitoring and external Doppler examination alone (control group). The primary outcomes measured were the number of true positive vascular thrombotic events and the number of false positive events resulting in concern for flap compromise. Secondary outcomes measured included rates of flap take back and flap failure. A P-value of less than 0.5 was used to determine significant differences between groups.
Perioperative Flap Monitoring
RESULTS
Intraoperative free flap monitoring began following completion of both the venous and arterial anastomoses. If the Flow Coupler was used, the wire was connected and the device was turned on at this time. Any subsequent Flow Coupler Doppler signal loss alerted surgeons to clinically examine the flap. If the signal was lost intra-
From May 2012 to March 2014, 72 patients underwent 119 abdominal based free flaps for breast reconstruction following mastectomy at the University of Wisconsin. In the first 50 patients (85 consecutive flaps), the Flow Coupler was used for the venous anastomosis and as an adjunct to clinical monitoring and external Doppler
Microsurgery DOI 10.1002/micr
Free Flap Monitoring Using Flow Coupler
339
Table 1. Demographic Information and Free Flap Data for the Flow Coupler (Group 1) and Non-Flow Coupler (Group 2) Patients
Number of patients Average age (years) Average BMI (kg/m2) Prior radiation Number of flaps DIEP MS TRAM SIEA Bilateral cases Intraoperative vessel thrombosis Postoperative vessel thrombosis Combined intra/postoperative vessel thrombosis Flap take back for flap compromise Flap take back salvage Flap failure Hospital stay (days) Device failure Wire pulled out Flow Coupler wire kinking vein
P-value (t-test or Fischer exact test)
Direct flap monitoring: Flow Coupler
Non-flow Coupler with clinical examination
Both groups combined
50 49.7 28.5 15 (17.6%) 85 31 (36.5%) 53 (62.3%) 1 (1.2%) 35 (70%) 9 (10.6%)
22 49.6 29.8 6 (17.6%) 34 9 (26.5%) 25 (73.5%) 0 (0%) 12 (55%) 0 (0%)
72 49.7 28.9 17.6% 119 33.6% 65.5% 0.8% 65.3% 7.5%
P 5 0.391 P 5 0.29 P 5 1.0 P 5 0.679 P 5 0.058
6 (7.9%)
1 (2.9%)
6.4%
P 5 0.43
15 (17.6%)
1 (2.9%)
13.4%
P 5 0.038
6 (7.1%)
0 (0%)
5%
P50.714
6 (100%) 4 (4.7%) 5.54 1 (1.2%) 6 (7.1%) 4 (4.7%)
NA 1 (2.9%) 4.97 NA NA NA
4.2% 5.4
NA P 5 1.0 P 5 0.130
examination (group 1). In the following 22 patients (34 consecutive flaps), the non-flow Coupler was used with clinical monitoring and external Doppler examination alone (group 2). The IMA and IMV were used in all cases. The average patient age was 49.7 years, average BMI was 28.9 kg/m2, and incidence of prior radiation was 17.6%. The overall incidence of vascular thrombotic events was 13.4% (17.6% in group 1; 2.9% in group 2; P 5 0.038). The rate of re-operation for suspected vascular compromise with intent for flap salvage was 5% (7.1% in group 1; 0% in group 2; P 5 0.181). Flap salvage rate in group 1 was 100%. The overall rate of flap failure was 4.2% (4.7% in group 1; 2.9% in group 2; P 5 1.0). None of the failed flaps were taken back to the operating room for salvage as they were either determined non-salvageable based on multiple attempts to improve flap perfusion at time of initial operation or patient condition did not permit timely return to the operating room. Average length of hospital stay in the groups 1 and 2 were 5.54 days and 4.97 days respectively (P 5 0.130) (Table 1). Intraoperative Flap Monitoring: Flow Coupler
In the intraoperative setting, Flow Coupler signal loss occurred in 25 flaps (29.4%). Both vessels were inspected
P 5 0.954 P 5 0.214 P 5 1.0
in all cases except for two flaps where the flap lost signal during cardiopulmonary resuscitation (CPR) at the conclusion of the case. A true vascular thrombotic problem was identified in nine flaps where three cases of arterial thrombosis, one venous thrombosis, and three cases with both arterial and venous thrombosis were identified. These nine thrombotic problems included the two flaps not explored. No vascular thrombotic problem was present in 16 of the intraoperative flap signal losses. Eleven flap signal losses returned with repositioning or re-inset of the flap. In four of these cases there was obvious kinking of the vein by the Flow Coupler wire. Four flap signal losses were due to inadvertent wire removal and one signal loss was secondary to device malfunction as there was normal venous and arterial flow (Fig. 2). In six of the 25 flaps that lost signal, signal loss occurred more than once requiring additional flap repositioning and re-inset. Postoperative Flap Monitoring: Flow Coupler
In the postoperative setting, 76 flaps were monitored using the Flow Coupler (9 flaps no longer monitored due to 4 intraoperative flap failures, 4 Flow Coupler wires removed, and 1 device malfunction). Signal loss occurred in 10 flaps (13.2%) postoperatively, with no flaps losing signal more than once. Six flaps returned to the operating Microsurgery DOI 10.1002/micr
340
Kempton et al.
Figure 2. Etiology of intraoperative free flap signal loss.
Figure 3. Etiology of postoperative free flap signal loss.
room due to clinical concern in which all six flaps required anastomotic revision due to thrombosis. The four flaps that lost signal and were not taken to the operating room all had a reassuring clinical examination and external Doppler examination, two of which had confirmed inadvertent wire removal (Fig. 3). In 48 flaps, the Flow Coupler wire position in relation to the IMA along with the location that the wire exited on the chest wall was recorded. In 38 flaps, the IMV was located medial to the artery with the Flow Coupler wire passing posterior to the artery, traversing under the flap before exiting the chest laterally. Six flaps with this configuration (15.8%) suffered vascular compromise requiring revision. The IMV was located lateral to the artery in five flaps with the wire exiting the chest laterally and no flaps in this group required revision. In five flaps, the IMV was located medial to the artery with the wire exiting the chest medially and one flap (20%) required anastomotic revision (Fig. 4). Flap arterial and venous configuration was not found to significantly relate to flap vascular compromise (P 5 1.0). Microsurgery DOI 10.1002/micr
Perioperative Flap Monitoring: Non-Flow Coupler
Abdominal-based free flaps for breast reconstruction using the non-flow Coupler (group 2) served as the control group. Intraoperatively, there were two flaps that appeared to be congested upon clinical assessment. One flap was explored and found to have venous insufficiency despite a patent venous anastomosis and strong arterial external Doppler examination. This flap was monitored closely postoperatively and failed on postoperative day 3 requiring debridement and another free flap from the left hemiabdomen. The other flap that appeared congested was found to have poor flow through the vein despite absence of thrombosis and a vein graft was used to anastomose the superficial epigastric vein to the IMV and the flap did well. There were no take backs for flap compromise. Comparisons of Sensitivity and Positive Predictive Value
Group 1 flaps using the Flow Coupler demonstrated 100% sensitivity in the intra- and postoperative settings.
Free Flap Monitoring Using Flow Coupler
341
PPV. Postoperative sensitivity and PPV could not be calculated for group 2 due to no true positive flap vascular compromise (Table 2). To assess the potential for a learning curve associated with using the Flow Coupler device, the Flow Coupler patients were separated into two groups (first 43 flaps: last 42 flaps) for comparison. There was no significant differences found in intra- or postoperative signal loss, false positive results, signal loss during inset, wire inadvertent removal, venous kinking, true vascular thrombotic events, flap take back, or flap failure between these flaps (Table 3). The last 42 Flow Coupler flaps were then compared to the 34 flaps performed with the use of the non-flow Coupler with clinical monitoring. In this comparison, there were significantly more true vascular thrombotic events with the use of the Flow Coupler (26.2% vs. 2.9%; P 5 0.0091). Additionally, the Flow Coupler resulted in significantly more false positive concern for flap compromise than the non-flow Coupler with clinical monitoring and external Doppler examination alone in the intraoperative setting (P 5 0.015) (Table 4). DISCUSSION
Figure 4. Observed relationships between the IMA and IMV in relation to the path of the Flow Coupler wire: (top) vein medial to artery with wire exiting the chest laterally; (middle) vein lateral to artery with wire exiting the chest laterally; (bottom) vein medial to artery with wire exiting the chest medially.
The intra- and postoperative PPV was 36% and 60% respectively (incidence that the Flow Coupler correctly identified a problem). For group 2, the intraoperative sensitivity of clinical examination was 100% with 100%
The Flow Coupler was shown in 85 consecutive autologous free flap breast reconstruction cases to be 100% sensitive in detecting flap compromise in both the intraand postoperative settings. However, the use of the Flow Coupler resulted in a high incidence of true vascular thrombotic events (17.6%), which was found to be significant when compared to use of the non-flow Coupler and resulted in the need for more anastomotic revisions. Additionally, the Flow Coupler had a high number of false positive events, with vascular compromise being present in 36% of intraoperative signal losses and 60% of postoperative signal losses. The number of false positive events was higher in the Flow Coupler group compared to the non-flow Coupler group with clinical monitoring alone, resulting in the need for significantly more intraoperative flap exploration and flap repositioning. The rates of flap take back and flap failure were higher in the Flow Coupler group, but this was not found to be statistically significant. A prominent criticism of venous implantable Doppler monitoring is the inability to distinguish between vascular thrombosis and technical device malfunction and the concern for the device causing kinking of the vessel.7,10,12,15,16 The reported frequency of postoperative false positive Doppler signal loss with the use of the Cook–Swartz system ranges from 0.9% to 35%.1,3,11,12,16–19 The occurrence of a false positive finding often complicates decision making for flap take back and can result in unnecessary trips to the operating room.1 In 20 free flaps for head and neck reconstruction Microsurgery DOI 10.1002/micr
342
Kempton et al. Table 2. Evaluation of Flow Coupler vs. Non-Flow Coupler Device Reliability Direct flap monitoring: Flow Coupler
Viable Flap True negative False positive Compromised flap True positive False negative PPV NPV Sensitivity Specificity
Non-flow Coupler with clinical examination
Intra-OP
Post-OP
Intra-OP
Post-OP
76 60 16 9 9 0 36% 100% 100% 79%
70 66 4 6 6 0 60% 100% 100% 94%
32 32 0 2 2 0 100% 100% 100% 100%
33 33 0 0 0 0 Undefined 100% Undefined 100%
Table 3. Comparison of First Flow Coupler Patients to Last Flow Coupler Patients
Intraoperatively Signal loss or clinical examination concern True positive False positive Inset Wire out Vein twist Vessel thrombosis Take back for vascular compromise Flap failure
Postoperatively Signal loss or clinical examination concern True positive False positive Inset Wire out Vein twist Vessel thrombosis Take back for vascular compromise Flap failure Intraoperative and postoperative vessel thrombosis
Flow Coupler (First 43)
Flow Coupler (Last 42)
P-value
12
13
P 5 0.815
3 8 4 4 1 3 NA
6 8 3 0 3 6 NA
P 5 0.313 P 5 1.00 P 5 1.00 P 5 0.116 P 5 0.36 P 5 0.313 NA
2
2
P 5 1.00
Flow Coupler (Last 40)
Flow Coupler (n 5 36)
P-value
4
6
P 5 0.503
1 3 NA 1 0 1 1
5 1 NA 1 0 5 5
P 5 0.095 P 5 0.617 NA P 5 1.00 P 5 1.00 P 5 0.095 P 5 0.095
0 4
0 11
P 5 1.00 P 5 0.050
by Zhang et al., the Flow Coupler was found to have two (10%) cases of postoperative false positive signal loss, in which neither flap was explored due to reassuring Microsurgery DOI 10.1002/micr
clinical examination.15 In our series, there were four cases (5.3%) of postoperative false positive signal loss where the clinical examination was reassuring and the flap did well without a take back to the operating room. Sole reliance on the venous flow coupler may have resulted in unnecessary flap exploration. Use of the Flow Coupler was most problematic in the intraoperative setting. This was unexpected as La Torre et al. commented on the advantage of intraoperative monitoring allowing for continuous feedback during flap inset.19 This advantage was appreciated in our series when the Flow Coupler signal was audible, as this was always associated with a viable free flap (100% sensitivity) and was particularly useful when bilateral free flaps were being performed (70% of cases). Loss of an audible Flow Coupler signal, however, was only associated with vessel thrombosis 36% of the time, resulting in unnecessary intraoperative maneuvers to confirm anastomotic patency, likely adding time to an already lengthy procedure. Flow Coupler device malfunction occurred in seven flaps in our series. In one case of signal loss, there was good flow across the vein despite the coupler and wire being in good position. In the other six cases, the wire was inadvertently pulled from the coupler secondary to patient agitation, coughing, or traction from patient positioning. In all seven cases, the flaps were monitored with clinical examination and did well. Ultimately, the absence of vessel thrombosis was used to define a false positive signal loss in our series. In 11 cases of intraoperative signal loss, however, the Flow Coupler signal returned with flap and/or anastomotic repositioning and in four of these cases, traction from the Flow Coupler wire was found to be the cause of twisting or stretching of the vein. While only speculative, the remaining seven flaps may have also had impending vascular compromise that corrected following release of inset sutures but prior to direct visualization of the anastomosis. In this case, a false positive may have become a true positive without intraoperative intervention. Therefore, we still recommend direct assessment of the anastomosed vessels in any case of intraoperative signal loss. True positive vascular thrombotic events were significantly higher in the Flow Coupler group (17.6%) when compared to the non-flow Coupler group (2.9%). It is unlikely that this high vascular thrombotic rate is due to the Coupler itself as the Flow Coupler is very similar to the non-flow Coupler, which has been shown to have a venous thrombosis rate of 0.6% in a series of 1,000 consecutive venous anastomoses in free flap breast reconstruction cases.14 Therefore, we speculate that this thrombosis rate is due to kinking or stretching of the venous anastomosis as a direct result of traction from the Flow Coupler wire. The orientation of most breast free
Free Flap Monitoring Using Flow Coupler Table 4. Comparison of Last Flow Coupler Patients to non-flow Coupler Patients
Intraoperatively Signal loss or clinical exam concern True positive False positive Inset wire out Vein twist Vessel Thrombosis Take back for vascular compromise Flap failure
Postoperatively Signal loss or clinical exam concern True positive False positive Inset Wire out Vein twist Vessel thrombosis Take back for vascular compromise Flap failure Total intraoperative and postoperative vessel thrombosis
Direct flap monitoring: Flow Coupler (Last 42)
Non-flow Coupler with clinical examination (n 5 34)
P-value
13
2
P 5 0.008
6 7 3 0 3 6 NA
2 0 2 NA 0 0 NA
P 5 0.285 P 5 0.015 P 5 1.00 NA P 5 0.248 P 5 0.030 NA
2
0
P 5 0.50
Direct flap monitoring: Flow Coupler (Last 36)
Non-flow Coupler with clinical examination (n 5 33)
P-value
6
0
P 5 0.025
5 1 NA 1 0 5 5
0 0 NA NA 0 1 0
P 5 0.055 P 5 1.00 NA NA P 5 1.00 P 5 0.19 P 5 0.055
0 11
1 1
P 5 0.48 P 5 0.0091
flaps is peculiar in that the flap is lying on top of the vessels, making it practically impossible to visualize the final position of the vessels following flap inset. In this orientation, the vessels may be more likely to kink if their mobility is limited by the presence of a fixed wire. A fixed wire position causing venous kinking is a commonly reported issue with the Cook-Swartz device.7,12,16 In 48 flaps where the IMA and IMV position were recorded in relation to the course of the Flow Coupler wire, vessel configuration and wire exit point were not found to contribute to vessel kinking or thrombosis. If the Flow Coupler is used, however, we feel that having the wire exit the chest wall skin on the side that allows the vein to sit in a natural position is important and that this importance may have been uncovered with a larger sample size.
343
Due to the results that are detailed in this article (e.g., high number of false positive signal losses and true vascular thrombotic events) the Flow Coupler was felt to be no longer useful in our breast reconstruction practice. It should be noted that this study focused on free flaps for breast reconstruction where the internal mammary vessels were used, and the results may not apply to other anatomical areas. Therefore, it is important that the reader does not generalize the results of this paper on breast reconstruction to other areas of the body. Our experience with the use of the Flow Coupler in head and neck and in lower extremity reconstruction cases has been minimal, however, we have, anecdotally, not had any false positive results or cases where the vessels were kinked or compressed by the vein. Of note, Zhang et al. in a series of 20 flaps monitored using the venous flow coupler, reported two cases (10%) of venous kinking caused by the wire.15 While there are no published reports on the use of the Flow Coupler in lower extremity free flap monitoring, Rozen et al. reported on 20 lower extremity free flaps monitored with the Cook–Swartz device and had no increase in flap compromise or false positive signal loss.17 This study had several limitations that are inherent to most case series. Although the comparison groups in this study are small, significance was found in the primary outcomes of interest (false positive and true positive events) and there is a 1-to-1 relationship between P-value and power. As studies are typically powered for a single primary outcome of interest, this study may have been underpowered to find differences in secondary outcomes including flap take back or flap failure. When looking at retrospective data for which a new device is used for a period of time and then not used for a period of time, the results are potentially impacted by a time effect and patient selection bias. Selection bias was minimal as patients were well matched for age, BMI, and prior radiation exposure. Since patient outcomes were only reviewed 2 weeks postoperatively, a time effect did not impact the results. Additionally, there may be a learning curve associated with the use of a new device. This potential was minimized as 25 free flaps were performed using the Flow Coupler prior to the prospective collection of data on patients receiving abdominal-based free flap breast reconstruction. Additionally, we found no significant difference in intra- or postoperative events between the first 43 and subsequent 42 venous flow coupler flaps. CONCLUSION
If the Flow Coupler Doppler signal is audible, it becomes an extremely sensitive method to confirm patency of a microsurgical anastomosis. However, there is a high false positive rate (loss of Doppler signal, no thrombotic problem) in both the intraoperative and postoperative settings. This led to the need for frequent Microsurgery DOI 10.1002/micr
344
Kempton et al.
intraoperative maneuvers to diagnose and amend the cause of signal loss. The Flow Coupler also resulted in significantly more true vascular thrombotic events when compared to the non-flow Coupler requiring the need for more vascular anastomotic revisions.
REFERENCES 1. Whitaker IS, Rozen WM, Chubb D, Acosta R, Kill BJ, BirkeSorensen H, Grinsell D, Ashton MW. Postoperative monitoring of free flaps in autologous breast reconstruction: A multicenter comparison of 398 flaps using clinical monitoring, microdialysis, and the implantable Doppler probe. J Reconstr Microsurg 2010;26:409–416. 2. Abdel-Galil K, Mitchell D. Postoperative monitoring of microsurgical free tissue transfers for head and neck reconstruction: A systematic review of current techniques-Part I. Non-invasive techniques. Br J Oral Maxillofac Surg 2009;47:351–355. 3. Kind GM, Buntic RF, Buncke GM, et al. The effect of an implantable Doppler probe on the salvage of microvascular tissue transplants. Plast Reconstr Surg 1998;101:1268–1273. 4. Luu Q, Farwell DG. Advances in free flap monitoring: Have we gone too far? Curr Opin Otolaryngol Head Neck Surg 2009;17: 267–269. 5. Fox PM, Zeidler K, Carey J, Lee GK. White light spectroscopy for free flap monitoring. Microsurgery 2013;33:198–202. 6. Rothenberger J, Amr A, Schaller HE, Rahmanian-Schwarz A. Evaluation of a non-invasive monitoring method for free flap breast reconstruction using laser Doppler flowmetric and tissue spectrophotometry. Microsurgery 2013;33:350–357. 7. Lohman R, Langevin CJ, Bozkurt M, Kundu N, Djohan R. A prospective analysis of free flap monitoring techniques: Physical examination, external Doppler, implantable Doppler, and tissue oxyimetry. J Recontsr Microsurg 2013;29:51–56. 8. Kahn MAA, Mohan A, Ahmed W, Rayatt S. Nursing monitoring and management of free flaps and pedicled flaps—Outcomes of teaching sessions on flap care. Plast Surg Nurs 2010;30:213–216.
Microsurgery DOI 10.1002/micr
9. Creech BJ, Miller SH. Evaluation of circulation in skin flaps. In: Skin Flaps. Grabb WC, Myers MB. (Eds) Boston: Little, Brown & Company; 1975. 10. Swartz WM, Jones NF, Cherup L, Klein A. Direct monitoring of microvascular anastomosis with the 20-mHz ultrasonic Doppler probe: An experimental and clinical study. Plast Reconstr Surg 1988;81:149–161. 11. Swartz WM, Izquierdo R, Miller MJ. Implantable venous Doppler microvascular monitoring: Laboratory investigation and clinical results. Plast Reconstr Surg 1994;93:152–163. 12. Paydar KZ, Hansen SL, Chang DS, Hoffman WY, Leon P. Implantable venous doppler monitoring in head and neck free flap reconstruction increases the salvage rate. Plast Reconstr Surg 2010;125: 1129–1134. 13. Rozen WM, Chubb D, Whitaker IS, Acosta R. The efficacy of postoperative monitoring: A single surgeon comparison of clinical monitoring and the implantable Doppler probe in 547 consecutive free flaps. Microsurgery 2010;30:105–110. 14. Jandall S, Wu LC, Vega SJ, Kovach SJ, Serletti JM. 1000 consecutive venous anastomoses using the microvascular anastomotic coupler in breast reconstruction. Plast Reconstr Surg 2010;125:792–798. 15. Zhang T, Dyalram-Silverberg D, Bui T, Caccamese JF Jr, Lubek JE. Analysis of an implantable venous anastomotic flow coupler: Experience in head and neck free flap reconstruction. Int J Oral Maxillofac Surg 2012;41:751–755. 16. Rosenberg JJ, Fornage BD, Chevray PM. Monitoring buried free flaps: Limitations of the implantable Doppler and use of color duplex sonography as a confirmatory test. Plast Reconstr Surg 2006; 118:109–113. 17. Rozen WM, Enajat M, Whitaker IS, Lindkvist U, Audolfsson T, Acosta R. Postoperative monitoring of lower limb free flaps with the Cook-Swartz implantable Doppler probe: A clinical trial. Microsurgery 2010;30:354–360. 18. Ferguson REH, Yu P. Techniques of monitoring buried fasciocutaneous free flaps. Plast Reconstr Surg 2009;123:525–532. 19. de la Torre J, Hedden W, Grant JH, Gardner PM, Fix RJ, Vasconez LO. Retrospective review of the internal doppler probe for intraand postoperative microvascular surveillance. J Reconstr Microsurg 2003;19:287–289.
Purpose: The purpose of this study is to evaluate the use of the venous anastomotic Flow Coupler in monitoring free flaps used for breast reconstruction in a consecutive series of patients. Methods: Retrospective data were collected on patients undergoing free flap breast reconstruction from May 2012 to March 2014. The venous anastomotic Flow Coupler was used in the first 85 flaps and a non-flow Coupler with clinical and external Doppler monitoring alone in the subsequent 34 flaps. Data collected included patient age, BMI, prior radiation, flap type, intra- and postoperative Flow Coupler events, along with rates of flap take back, salvage, and failure. Proportion data were compiled and statistically analyzed. Results: One hundred nineteen consecutive abdominal based breast reconstruction free flaps were performed. The overall flap failure rate was 4.2% (4.7% Flow Coupler; 2.9% in non-flow Coupler; P 5 1.0). The Flow Coupler demonstrated 100% sensitivity in the intra- and postoperative settings. A positive predictive value of 36% was noted intraoperatively which was significantly higher compared to the non-flow Coupler group (P 5 0.015). Vessel thrombosis occurred in 17.6% of Flow Coupler flaps, which was significantly higher when compared to the non-flow Coupler (2.9%; P 5 0.038). Conclusions: The Flow Coupler is a sensitive method to confirm patency of a microsurgical anastomosis. However, there is a high false positive rate in both the intraoperative and postoperative settings resulting in frequent intraoperative maneuvers to amend the cause of signal loss. Additionally, the Flow Coupler resulted C 2014 Wiley Periodicals, Inc. Microsurgery in significantly more vascular thrombotic events when compared to the non-flow Coupler. V 35:337–344, 2015.
There
is substantial variation in flap monitoring techniques following free tissue transfer including surface temperature recording, transcutaneous laser and pencil Doppler, PO2 monitoring, visible white light spectroscopy, intravenous fluorescein, photoplethysmography, and implantable Doppler monitoring.1–7 Clinical observation remains the gold standard assessment of flap viability, but is labor intensive and requires skill and experience.7,8 The ideal device should be safe for the patient and the flap, allow for continuous monitoring, be sensitive and reliable in diagnosing both venous and arterial insufficiency, and be applicable to all flap types.9 In the 1980s, The Cook–Swartz implantable Doppler system was the first to enable direct anastomotic Doppler monitoring, consisting of a Doppler probe attached near the venous anastomosis using a silicone leash.10 This method of flap monitoring has demonstrated reliability is multiple series, however, requires additional operative steps and can be difficult to apply.1,3,11–13 The advent of the venous anastomotic Coupler by Synovis Life Technologies, Inc. has changed the way the venous anastomosis is performed and Division of Plastic and Reconstructive Surgery, University of Wisconsin Hospital and Clinics, Madison, WI Grant sponsor: Division of Plastic and Reconstructive Surgery at University of Wisconsin Hospital and Clinics. *Correspondence to: Ahmed M. Afifi, M.D., Division of Plastic and Reconstructive Surgery, Department of Surgery, University of Wisconsin Hospital and Clinics, 600 Highland Avenue, G5/361 Clinical Science Center, Mail code 3236, Madison, WI 53792. E-mail: [email protected] Received 3 June 2014; Revision accepted 26 September 2014; Accepted 3 October 2014 Published online 21 October 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/micr.22341 Ó 2014 Wiley Periodicals, Inc.
has been shown to be both safe and effective when compared to hand sewn techniques.14 In 2010, the Flow Coupler was introduced by Synovis Life Technologies, Inc., presenting a fusion of the venous anastomotic Coupler and a 20 MHz ultrasonic microdoppler probe. While the Flow Coupler is being used in clinic microsurgical practice, published date on the use of this device is limited.15 The aim of this work is to evaluate the use of the Flow Coupler in monitoring free flaps used for breast reconstruction in a consecutive series of patients, and describe intra- and postoperative events. PATIENTS AND METHODS
In December of 2011, the two senior authors (AA, SP) initiated the use of the venous anastomotic Flow Coupler to serve as an adjunct to clinical and external Doppler free flap monitoring (Fig. 1) (Synovis Micro Companies Alliance Inc., Birmingham, AL). From May of 2013, a Synovis Life Technologies, Inc. Coupler without an implantable Doppler probe (non-flow Coupler) was used in all subsequent abdominal-based free flaps. In these patients, external (pencil) Doppler examination and clinical assessment alone was used for free flap monitoring. Data included the occurrence, cause, and management of all intraoperative or postoperative events such as signal loss or flap vascular compromise. Following Institutional Review Board approval, hospital records, operative reports, and prospectively collected perioperative flap monitoring events were reviewed. Demographic data collected included patient age, BMI, and prior radiation. Perioperative data included type of
338
Kempton et al.
Figure 1. Intraoperative photo of venous anastomotic flow coupler with wire passing under the artery to exit the chest wall laterally.
abdominal-based flap (muscle sparing transverse rectus abdominis myocutaneous [MS-TRAM], deep inferior epigastric perforator [DIEP], superifical inferior epigastric artery [SIEA]), recipient vessel used for anastomosis, intraoperative and postoperative Flow Coupler events, along with rate of flap take back, salvage, and failure. After the first event of venous kinking that was found to be caused by traction from the Flow Coupler wire, the location of the internal mammary vein (IMV) in relation to the internal mammary artery (IMA), the course of the Flow Coupler wire, and wire exit point (medial vs. lateral) on the chest wall was additionally recorded. Application of Venous Anastomotic Flow Coupler
Abdominal-based free flap harvest and recipient vessel preparation was done simultaneously. The preferred recipient vessels were the IMA and IMV. Venous anastomosis always preceded arterial anastomosis. The recipient and donor veins were clamped separately. A vesselmeasuring gauge was used to determine Flow Coupler size. Donor and recipient veins were anastomosed using the Flow Coupler in the usual fashion. The arterial anastomosis was then completed using interrupted 8-0 nylon sutures. The Doppler wire was passed out percutaneously through a stab incision close to the IMF medially or laterally. In the setting of bilateral reconstruction, the above was repeated on the contralateral side.
operatively, the flap vessels and the anastomosis were examined, often requiring removal of the flap insetting sutures (between the flap and the chest wall). If there was good flow in the vessels, the flap was repositioned and the sutures used for inset were replaced. Thrombotic and occlusion problems were dealt with accordingly. The Doppler signals were continuously monitored following anastomosis until the patient left the operating room. Patients were transferred to the regular unit where flaps were monitored every hour for the first 24 hours, every 2 hours for the next 48 hours, then every 4 hours until discharge. Flap monitoring consisted of clinical examination (capillary refill, temperature, blanching, and skin turgor), Flow Coupler signal monitoring if the Flow Coupler was used, and external Doppler examination. Operative take back was defined as return to the operating room for flap vascular compromise and therefore did not include evacuation of hematoma in the setting of a viable flap, debridement of non-viable mastectomy skin flaps, or any other secondary procedures. Data Analysis
Flap monitoring during the intraoperative and postoperative periods were analyzed separately. Only one signal loss in the intraoperative and one signal loss in the postoperative setting were used for the analysis. A true positive event was defined as loss of signal and a compromised flap. A false positive event was the loss of a signal in the setting of a viable flap. A true negative event was defined as no loss of signal and a viable flap. A false negative event was no loss of signal and a compromised flap. Positive and negative predictive values (PPV/NPV), sensitivity, and specificity were calculated to describe the performance of the Flow Coupler. Student’s t-test and Fischer exact test were used to compare data between breast reconstruction cases with use of the Flow Coupler and cases that used the nonflow Coupler with clinical monitoring and external Doppler examination alone (control group). The primary outcomes measured were the number of true positive vascular thrombotic events and the number of false positive events resulting in concern for flap compromise. Secondary outcomes measured included rates of flap take back and flap failure. A P-value of less than 0.5 was used to determine significant differences between groups.
Perioperative Flap Monitoring
RESULTS
Intraoperative free flap monitoring began following completion of both the venous and arterial anastomoses. If the Flow Coupler was used, the wire was connected and the device was turned on at this time. Any subsequent Flow Coupler Doppler signal loss alerted surgeons to clinically examine the flap. If the signal was lost intra-
From May 2012 to March 2014, 72 patients underwent 119 abdominal based free flaps for breast reconstruction following mastectomy at the University of Wisconsin. In the first 50 patients (85 consecutive flaps), the Flow Coupler was used for the venous anastomosis and as an adjunct to clinical monitoring and external Doppler
Microsurgery DOI 10.1002/micr
Free Flap Monitoring Using Flow Coupler
339
Table 1. Demographic Information and Free Flap Data for the Flow Coupler (Group 1) and Non-Flow Coupler (Group 2) Patients
Number of patients Average age (years) Average BMI (kg/m2) Prior radiation Number of flaps DIEP MS TRAM SIEA Bilateral cases Intraoperative vessel thrombosis Postoperative vessel thrombosis Combined intra/postoperative vessel thrombosis Flap take back for flap compromise Flap take back salvage Flap failure Hospital stay (days) Device failure Wire pulled out Flow Coupler wire kinking vein
P-value (t-test or Fischer exact test)
Direct flap monitoring: Flow Coupler
Non-flow Coupler with clinical examination
Both groups combined
50 49.7 28.5 15 (17.6%) 85 31 (36.5%) 53 (62.3%) 1 (1.2%) 35 (70%) 9 (10.6%)
22 49.6 29.8 6 (17.6%) 34 9 (26.5%) 25 (73.5%) 0 (0%) 12 (55%) 0 (0%)
72 49.7 28.9 17.6% 119 33.6% 65.5% 0.8% 65.3% 7.5%
P 5 0.391 P 5 0.29 P 5 1.0 P 5 0.679 P 5 0.058
6 (7.9%)
1 (2.9%)
6.4%
P 5 0.43
15 (17.6%)
1 (2.9%)
13.4%
P 5 0.038
6 (7.1%)
0 (0%)
5%
P50.714
6 (100%) 4 (4.7%) 5.54 1 (1.2%) 6 (7.1%) 4 (4.7%)
NA 1 (2.9%) 4.97 NA NA NA
4.2% 5.4
NA P 5 1.0 P 5 0.130
examination (group 1). In the following 22 patients (34 consecutive flaps), the non-flow Coupler was used with clinical monitoring and external Doppler examination alone (group 2). The IMA and IMV were used in all cases. The average patient age was 49.7 years, average BMI was 28.9 kg/m2, and incidence of prior radiation was 17.6%. The overall incidence of vascular thrombotic events was 13.4% (17.6% in group 1; 2.9% in group 2; P 5 0.038). The rate of re-operation for suspected vascular compromise with intent for flap salvage was 5% (7.1% in group 1; 0% in group 2; P 5 0.181). Flap salvage rate in group 1 was 100%. The overall rate of flap failure was 4.2% (4.7% in group 1; 2.9% in group 2; P 5 1.0). None of the failed flaps were taken back to the operating room for salvage as they were either determined non-salvageable based on multiple attempts to improve flap perfusion at time of initial operation or patient condition did not permit timely return to the operating room. Average length of hospital stay in the groups 1 and 2 were 5.54 days and 4.97 days respectively (P 5 0.130) (Table 1). Intraoperative Flap Monitoring: Flow Coupler
In the intraoperative setting, Flow Coupler signal loss occurred in 25 flaps (29.4%). Both vessels were inspected
P 5 0.954 P 5 0.214 P 5 1.0
in all cases except for two flaps where the flap lost signal during cardiopulmonary resuscitation (CPR) at the conclusion of the case. A true vascular thrombotic problem was identified in nine flaps where three cases of arterial thrombosis, one venous thrombosis, and three cases with both arterial and venous thrombosis were identified. These nine thrombotic problems included the two flaps not explored. No vascular thrombotic problem was present in 16 of the intraoperative flap signal losses. Eleven flap signal losses returned with repositioning or re-inset of the flap. In four of these cases there was obvious kinking of the vein by the Flow Coupler wire. Four flap signal losses were due to inadvertent wire removal and one signal loss was secondary to device malfunction as there was normal venous and arterial flow (Fig. 2). In six of the 25 flaps that lost signal, signal loss occurred more than once requiring additional flap repositioning and re-inset. Postoperative Flap Monitoring: Flow Coupler
In the postoperative setting, 76 flaps were monitored using the Flow Coupler (9 flaps no longer monitored due to 4 intraoperative flap failures, 4 Flow Coupler wires removed, and 1 device malfunction). Signal loss occurred in 10 flaps (13.2%) postoperatively, with no flaps losing signal more than once. Six flaps returned to the operating Microsurgery DOI 10.1002/micr
340
Kempton et al.
Figure 2. Etiology of intraoperative free flap signal loss.
Figure 3. Etiology of postoperative free flap signal loss.
room due to clinical concern in which all six flaps required anastomotic revision due to thrombosis. The four flaps that lost signal and were not taken to the operating room all had a reassuring clinical examination and external Doppler examination, two of which had confirmed inadvertent wire removal (Fig. 3). In 48 flaps, the Flow Coupler wire position in relation to the IMA along with the location that the wire exited on the chest wall was recorded. In 38 flaps, the IMV was located medial to the artery with the Flow Coupler wire passing posterior to the artery, traversing under the flap before exiting the chest laterally. Six flaps with this configuration (15.8%) suffered vascular compromise requiring revision. The IMV was located lateral to the artery in five flaps with the wire exiting the chest laterally and no flaps in this group required revision. In five flaps, the IMV was located medial to the artery with the wire exiting the chest medially and one flap (20%) required anastomotic revision (Fig. 4). Flap arterial and venous configuration was not found to significantly relate to flap vascular compromise (P 5 1.0). Microsurgery DOI 10.1002/micr
Perioperative Flap Monitoring: Non-Flow Coupler
Abdominal-based free flaps for breast reconstruction using the non-flow Coupler (group 2) served as the control group. Intraoperatively, there were two flaps that appeared to be congested upon clinical assessment. One flap was explored and found to have venous insufficiency despite a patent venous anastomosis and strong arterial external Doppler examination. This flap was monitored closely postoperatively and failed on postoperative day 3 requiring debridement and another free flap from the left hemiabdomen. The other flap that appeared congested was found to have poor flow through the vein despite absence of thrombosis and a vein graft was used to anastomose the superficial epigastric vein to the IMV and the flap did well. There were no take backs for flap compromise. Comparisons of Sensitivity and Positive Predictive Value
Group 1 flaps using the Flow Coupler demonstrated 100% sensitivity in the intra- and postoperative settings.
Free Flap Monitoring Using Flow Coupler
341
PPV. Postoperative sensitivity and PPV could not be calculated for group 2 due to no true positive flap vascular compromise (Table 2). To assess the potential for a learning curve associated with using the Flow Coupler device, the Flow Coupler patients were separated into two groups (first 43 flaps: last 42 flaps) for comparison. There was no significant differences found in intra- or postoperative signal loss, false positive results, signal loss during inset, wire inadvertent removal, venous kinking, true vascular thrombotic events, flap take back, or flap failure between these flaps (Table 3). The last 42 Flow Coupler flaps were then compared to the 34 flaps performed with the use of the non-flow Coupler with clinical monitoring. In this comparison, there were significantly more true vascular thrombotic events with the use of the Flow Coupler (26.2% vs. 2.9%; P 5 0.0091). Additionally, the Flow Coupler resulted in significantly more false positive concern for flap compromise than the non-flow Coupler with clinical monitoring and external Doppler examination alone in the intraoperative setting (P 5 0.015) (Table 4). DISCUSSION
Figure 4. Observed relationships between the IMA and IMV in relation to the path of the Flow Coupler wire: (top) vein medial to artery with wire exiting the chest laterally; (middle) vein lateral to artery with wire exiting the chest laterally; (bottom) vein medial to artery with wire exiting the chest medially.
The intra- and postoperative PPV was 36% and 60% respectively (incidence that the Flow Coupler correctly identified a problem). For group 2, the intraoperative sensitivity of clinical examination was 100% with 100%
The Flow Coupler was shown in 85 consecutive autologous free flap breast reconstruction cases to be 100% sensitive in detecting flap compromise in both the intraand postoperative settings. However, the use of the Flow Coupler resulted in a high incidence of true vascular thrombotic events (17.6%), which was found to be significant when compared to use of the non-flow Coupler and resulted in the need for more anastomotic revisions. Additionally, the Flow Coupler had a high number of false positive events, with vascular compromise being present in 36% of intraoperative signal losses and 60% of postoperative signal losses. The number of false positive events was higher in the Flow Coupler group compared to the non-flow Coupler group with clinical monitoring alone, resulting in the need for significantly more intraoperative flap exploration and flap repositioning. The rates of flap take back and flap failure were higher in the Flow Coupler group, but this was not found to be statistically significant. A prominent criticism of venous implantable Doppler monitoring is the inability to distinguish between vascular thrombosis and technical device malfunction and the concern for the device causing kinking of the vessel.7,10,12,15,16 The reported frequency of postoperative false positive Doppler signal loss with the use of the Cook–Swartz system ranges from 0.9% to 35%.1,3,11,12,16–19 The occurrence of a false positive finding often complicates decision making for flap take back and can result in unnecessary trips to the operating room.1 In 20 free flaps for head and neck reconstruction Microsurgery DOI 10.1002/micr
342
Kempton et al. Table 2. Evaluation of Flow Coupler vs. Non-Flow Coupler Device Reliability Direct flap monitoring: Flow Coupler
Viable Flap True negative False positive Compromised flap True positive False negative PPV NPV Sensitivity Specificity
Non-flow Coupler with clinical examination
Intra-OP
Post-OP
Intra-OP
Post-OP
76 60 16 9 9 0 36% 100% 100% 79%
70 66 4 6 6 0 60% 100% 100% 94%
32 32 0 2 2 0 100% 100% 100% 100%
33 33 0 0 0 0 Undefined 100% Undefined 100%
Table 3. Comparison of First Flow Coupler Patients to Last Flow Coupler Patients
Intraoperatively Signal loss or clinical examination concern True positive False positive Inset Wire out Vein twist Vessel thrombosis Take back for vascular compromise Flap failure
Postoperatively Signal loss or clinical examination concern True positive False positive Inset Wire out Vein twist Vessel thrombosis Take back for vascular compromise Flap failure Intraoperative and postoperative vessel thrombosis
Flow Coupler (First 43)
Flow Coupler (Last 42)
P-value
12
13
P 5 0.815
3 8 4 4 1 3 NA
6 8 3 0 3 6 NA
P 5 0.313 P 5 1.00 P 5 1.00 P 5 0.116 P 5 0.36 P 5 0.313 NA
2
2
P 5 1.00
Flow Coupler (Last 40)
Flow Coupler (n 5 36)
P-value
4
6
P 5 0.503
1 3 NA 1 0 1 1
5 1 NA 1 0 5 5
P 5 0.095 P 5 0.617 NA P 5 1.00 P 5 1.00 P 5 0.095 P 5 0.095
0 4
0 11
P 5 1.00 P 5 0.050
by Zhang et al., the Flow Coupler was found to have two (10%) cases of postoperative false positive signal loss, in which neither flap was explored due to reassuring Microsurgery DOI 10.1002/micr
clinical examination.15 In our series, there were four cases (5.3%) of postoperative false positive signal loss where the clinical examination was reassuring and the flap did well without a take back to the operating room. Sole reliance on the venous flow coupler may have resulted in unnecessary flap exploration. Use of the Flow Coupler was most problematic in the intraoperative setting. This was unexpected as La Torre et al. commented on the advantage of intraoperative monitoring allowing for continuous feedback during flap inset.19 This advantage was appreciated in our series when the Flow Coupler signal was audible, as this was always associated with a viable free flap (100% sensitivity) and was particularly useful when bilateral free flaps were being performed (70% of cases). Loss of an audible Flow Coupler signal, however, was only associated with vessel thrombosis 36% of the time, resulting in unnecessary intraoperative maneuvers to confirm anastomotic patency, likely adding time to an already lengthy procedure. Flow Coupler device malfunction occurred in seven flaps in our series. In one case of signal loss, there was good flow across the vein despite the coupler and wire being in good position. In the other six cases, the wire was inadvertently pulled from the coupler secondary to patient agitation, coughing, or traction from patient positioning. In all seven cases, the flaps were monitored with clinical examination and did well. Ultimately, the absence of vessel thrombosis was used to define a false positive signal loss in our series. In 11 cases of intraoperative signal loss, however, the Flow Coupler signal returned with flap and/or anastomotic repositioning and in four of these cases, traction from the Flow Coupler wire was found to be the cause of twisting or stretching of the vein. While only speculative, the remaining seven flaps may have also had impending vascular compromise that corrected following release of inset sutures but prior to direct visualization of the anastomosis. In this case, a false positive may have become a true positive without intraoperative intervention. Therefore, we still recommend direct assessment of the anastomosed vessels in any case of intraoperative signal loss. True positive vascular thrombotic events were significantly higher in the Flow Coupler group (17.6%) when compared to the non-flow Coupler group (2.9%). It is unlikely that this high vascular thrombotic rate is due to the Coupler itself as the Flow Coupler is very similar to the non-flow Coupler, which has been shown to have a venous thrombosis rate of 0.6% in a series of 1,000 consecutive venous anastomoses in free flap breast reconstruction cases.14 Therefore, we speculate that this thrombosis rate is due to kinking or stretching of the venous anastomosis as a direct result of traction from the Flow Coupler wire. The orientation of most breast free
Free Flap Monitoring Using Flow Coupler Table 4. Comparison of Last Flow Coupler Patients to non-flow Coupler Patients
Intraoperatively Signal loss or clinical exam concern True positive False positive Inset wire out Vein twist Vessel Thrombosis Take back for vascular compromise Flap failure
Postoperatively Signal loss or clinical exam concern True positive False positive Inset Wire out Vein twist Vessel thrombosis Take back for vascular compromise Flap failure Total intraoperative and postoperative vessel thrombosis
Direct flap monitoring: Flow Coupler (Last 42)
Non-flow Coupler with clinical examination (n 5 34)
P-value
13
2
P 5 0.008
6 7 3 0 3 6 NA
2 0 2 NA 0 0 NA
P 5 0.285 P 5 0.015 P 5 1.00 NA P 5 0.248 P 5 0.030 NA
2
0
P 5 0.50
Direct flap monitoring: Flow Coupler (Last 36)
Non-flow Coupler with clinical examination (n 5 33)
P-value
6
0
P 5 0.025
5 1 NA 1 0 5 5
0 0 NA NA 0 1 0
P 5 0.055 P 5 1.00 NA NA P 5 1.00 P 5 0.19 P 5 0.055
0 11
1 1
P 5 0.48 P 5 0.0091
flaps is peculiar in that the flap is lying on top of the vessels, making it practically impossible to visualize the final position of the vessels following flap inset. In this orientation, the vessels may be more likely to kink if their mobility is limited by the presence of a fixed wire. A fixed wire position causing venous kinking is a commonly reported issue with the Cook-Swartz device.7,12,16 In 48 flaps where the IMA and IMV position were recorded in relation to the course of the Flow Coupler wire, vessel configuration and wire exit point were not found to contribute to vessel kinking or thrombosis. If the Flow Coupler is used, however, we feel that having the wire exit the chest wall skin on the side that allows the vein to sit in a natural position is important and that this importance may have been uncovered with a larger sample size.
343
Due to the results that are detailed in this article (e.g., high number of false positive signal losses and true vascular thrombotic events) the Flow Coupler was felt to be no longer useful in our breast reconstruction practice. It should be noted that this study focused on free flaps for breast reconstruction where the internal mammary vessels were used, and the results may not apply to other anatomical areas. Therefore, it is important that the reader does not generalize the results of this paper on breast reconstruction to other areas of the body. Our experience with the use of the Flow Coupler in head and neck and in lower extremity reconstruction cases has been minimal, however, we have, anecdotally, not had any false positive results or cases where the vessels were kinked or compressed by the vein. Of note, Zhang et al. in a series of 20 flaps monitored using the venous flow coupler, reported two cases (10%) of venous kinking caused by the wire.15 While there are no published reports on the use of the Flow Coupler in lower extremity free flap monitoring, Rozen et al. reported on 20 lower extremity free flaps monitored with the Cook–Swartz device and had no increase in flap compromise or false positive signal loss.17 This study had several limitations that are inherent to most case series. Although the comparison groups in this study are small, significance was found in the primary outcomes of interest (false positive and true positive events) and there is a 1-to-1 relationship between P-value and power. As studies are typically powered for a single primary outcome of interest, this study may have been underpowered to find differences in secondary outcomes including flap take back or flap failure. When looking at retrospective data for which a new device is used for a period of time and then not used for a period of time, the results are potentially impacted by a time effect and patient selection bias. Selection bias was minimal as patients were well matched for age, BMI, and prior radiation exposure. Since patient outcomes were only reviewed 2 weeks postoperatively, a time effect did not impact the results. Additionally, there may be a learning curve associated with the use of a new device. This potential was minimized as 25 free flaps were performed using the Flow Coupler prior to the prospective collection of data on patients receiving abdominal-based free flap breast reconstruction. Additionally, we found no significant difference in intra- or postoperative events between the first 43 and subsequent 42 venous flow coupler flaps. CONCLUSION
If the Flow Coupler Doppler signal is audible, it becomes an extremely sensitive method to confirm patency of a microsurgical anastomosis. However, there is a high false positive rate (loss of Doppler signal, no thrombotic problem) in both the intraoperative and postoperative settings. This led to the need for frequent Microsurgery DOI 10.1002/micr
344
Kempton et al.
intraoperative maneuvers to diagnose and amend the cause of signal loss. The Flow Coupler also resulted in significantly more true vascular thrombotic events when compared to the non-flow Coupler requiring the need for more vascular anastomotic revisions.
REFERENCES 1. Whitaker IS, Rozen WM, Chubb D, Acosta R, Kill BJ, BirkeSorensen H, Grinsell D, Ashton MW. Postoperative monitoring of free flaps in autologous breast reconstruction: A multicenter comparison of 398 flaps using clinical monitoring, microdialysis, and the implantable Doppler probe. J Reconstr Microsurg 2010;26:409–416. 2. Abdel-Galil K, Mitchell D. Postoperative monitoring of microsurgical free tissue transfers for head and neck reconstruction: A systematic review of current techniques-Part I. Non-invasive techniques. Br J Oral Maxillofac Surg 2009;47:351–355. 3. Kind GM, Buntic RF, Buncke GM, et al. The effect of an implantable Doppler probe on the salvage of microvascular tissue transplants. Plast Reconstr Surg 1998;101:1268–1273. 4. Luu Q, Farwell DG. Advances in free flap monitoring: Have we gone too far? Curr Opin Otolaryngol Head Neck Surg 2009;17: 267–269. 5. Fox PM, Zeidler K, Carey J, Lee GK. White light spectroscopy for free flap monitoring. Microsurgery 2013;33:198–202. 6. Rothenberger J, Amr A, Schaller HE, Rahmanian-Schwarz A. Evaluation of a non-invasive monitoring method for free flap breast reconstruction using laser Doppler flowmetric and tissue spectrophotometry. Microsurgery 2013;33:350–357. 7. Lohman R, Langevin CJ, Bozkurt M, Kundu N, Djohan R. A prospective analysis of free flap monitoring techniques: Physical examination, external Doppler, implantable Doppler, and tissue oxyimetry. J Recontsr Microsurg 2013;29:51–56. 8. Kahn MAA, Mohan A, Ahmed W, Rayatt S. Nursing monitoring and management of free flaps and pedicled flaps—Outcomes of teaching sessions on flap care. Plast Surg Nurs 2010;30:213–216.
Microsurgery DOI 10.1002/micr
9. Creech BJ, Miller SH. Evaluation of circulation in skin flaps. In: Skin Flaps. Grabb WC, Myers MB. (Eds) Boston: Little, Brown & Company; 1975. 10. Swartz WM, Jones NF, Cherup L, Klein A. Direct monitoring of microvascular anastomosis with the 20-mHz ultrasonic Doppler probe: An experimental and clinical study. Plast Reconstr Surg 1988;81:149–161. 11. Swartz WM, Izquierdo R, Miller MJ. Implantable venous Doppler microvascular monitoring: Laboratory investigation and clinical results. Plast Reconstr Surg 1994;93:152–163. 12. Paydar KZ, Hansen SL, Chang DS, Hoffman WY, Leon P. Implantable venous doppler monitoring in head and neck free flap reconstruction increases the salvage rate. Plast Reconstr Surg 2010;125: 1129–1134. 13. Rozen WM, Chubb D, Whitaker IS, Acosta R. The efficacy of postoperative monitoring: A single surgeon comparison of clinical monitoring and the implantable Doppler probe in 547 consecutive free flaps. Microsurgery 2010;30:105–110. 14. Jandall S, Wu LC, Vega SJ, Kovach SJ, Serletti JM. 1000 consecutive venous anastomoses using the microvascular anastomotic coupler in breast reconstruction. Plast Reconstr Surg 2010;125:792–798. 15. Zhang T, Dyalram-Silverberg D, Bui T, Caccamese JF Jr, Lubek JE. Analysis of an implantable venous anastomotic flow coupler: Experience in head and neck free flap reconstruction. Int J Oral Maxillofac Surg 2012;41:751–755. 16. Rosenberg JJ, Fornage BD, Chevray PM. Monitoring buried free flaps: Limitations of the implantable Doppler and use of color duplex sonography as a confirmatory test. Plast Reconstr Surg 2006; 118:109–113. 17. Rozen WM, Enajat M, Whitaker IS, Lindkvist U, Audolfsson T, Acosta R. Postoperative monitoring of lower limb free flaps with the Cook-Swartz implantable Doppler probe: A clinical trial. Microsurgery 2010;30:354–360. 18. Ferguson REH, Yu P. Techniques of monitoring buried fasciocutaneous free flaps. Plast Reconstr Surg 2009;123:525–532. 19. de la Torre J, Hedden W, Grant JH, Gardner PM, Fix RJ, Vasconez LO. Retrospective review of the internal doppler probe for intraand postoperative microvascular surveillance. J Reconstr Microsurg 2003;19:287–289.
