The Implantable Cook-Swartz Doppler Probe for Postoperative Monitoring in Head and Neck Free Flap Reconstruction

Objective  To determine if the implantable Cook-Swartz Doppler Flow Monitoring System (Cook Vascular Inc, Vandergrift, Pennsylvania) improves surgical salvage rates for compromised free flaps.

Design  Retrospective medical record review spanning 2002 to 2006 for a large head and neck oncology program.

Setting  A tertiary care hospital.

Patients  A consecutive series of 351 patients (244 men and 107 women; mean age, 58.63 years) who underwent free flap reconstruction of head and neck defects that were monitored using the implantable Doppler probe were included.

Results  The most common indication for surgery was squamous cell carcinoma (81.0%), followed by functional reconstruction (4.3%). The most common free flap used was radial forearm (68.0%), followed by the fibular free flap (19.0%). With operative exploration used as the gold standard, the Cook-Swartz Doppler Flow Monitoring System had a sensitivity of 65.8% and specificity of 98.2% for the detection of flap compromise. For the detection of vascular compromise of the monitored vessel (excluding flap compromise cases whereby flow in the monitored vessel was not compromised on operative exploration, ie, venous obstruction, hematoma formation, and necrotizing fasciitis), the sensitivity increased to 100%.

Conclusions  This is the largest reported series, to our knowledge, of implantable Cook-Swartz Doppler use, and our experience would suggest that this is a reliable technique for postoperative monitoring in head and neck reconstruction. Our use of the implantable Doppler probe allowed us to recognize vascular compromise early, resulting in an overall flap success rate of 98.1%, with a 92.0% salvage rate of flaps that experienced vascular compromise of the monitored vessel.

Success rates of microvascular flap transplantation have increased dramatically with the continued refinement of microvascular techniques over the past 30 years, with many large studies citing success rates above 90%.^1-4 Conversely, salvage rates of free tissue transfers after vascular thrombosis or insufficiency have not increased to the same extent.^2,5,6 Timely reexploration and reanastomosis can salvage many failing free flaps; therefore, we must monitor each of the flaps very closely. At present, postoperative monitoring relies heavily on clinical evaluation using modalities such as temperature, skin color, capillary refill, bleeding to pinprick, and tissue turgor. Numerous devices have been introduced to monitor the patency of microvascular repairs, such as intravenous fluorescein,⁷ optical spectroscopy,⁸ the transcutaneous laser Doppler,⁹ photoplethysmography,¹⁰ and transcutaneous PO₂ monitoring.¹¹ However, each of these methods suffers from a major drawback in that they provide indirect evidence of vascular occlusion and therefore may not only be delayed in signaling vascular impairment but might not reflect what is actually happening at the anastomotic site.^12,13

Reliable postoperative monitoring is crucial in the detection of impaired flap perfusion and early detection may lead to successful salvage of the compromised flap. The ideal technique of monitoring skin flap viability would be rapidly responsive to vascular impairment and provide continuous, accurate information in a noninvasive manner.^3,6

In 1988, Swartz et al¹² introduced the technique of using silicone to secure a 1.0-mm Doppler probe to a cuff of expanded polytetrafluoroethylene (GORE-TEX; W. L. Gore and Associates, Flagstaff, Arizona), which was then secured around the flap vessel and sutured in place providing real-time monitoring of the flap blood flow. Recent studies using the implantable Doppler probe have shown improved detection of ischemia, particularly when monitoring buried flaps.^{1,3,5,6,13-15} In the present article, we review our experience with the use of the Cook-Swartz Doppler Flow Monitoring System (Cook Vascular Inc, Vandergrift, Pennsylvania) in a large series of head and neck free flap reconstructions. The objective of this study was to determine if this system improves the surgical salvage rates for compromised free flaps, as well as to calculate the sensitivity and specificity of this system for detecting compromise in the monitored vessel.

The ethics review board of the University of Alberta, Edmonton, Alberta, Canada, approved this study. A retrospective search of the 5-year database of a large head and neck oncology program was performed. As of December 2001, it became standard practice to use the Doppler monitoring system in all patients undergoing free flap reconstruction; therefore, all patients undergoing free flap reconstruction of head and neck defects between December 2001 and April 2006 were included in this study.

Collected information included patient demographics, indication for surgery, primary tumor site, free flap donor site(s), vessel(s) receiving Doppler probes, Doppler status, flap compromise, time to exploration, exploration findings, flap salvage, and flap failure or success. Operative exploration is currently the most sensitive and specific confirmatory test for flap compromise and was used as the gold standard for comparison in this study.

“Patient age” was noted at the time of free flap surgery. “Indication for surgery” was documented as the type of neoplasm, or for nonneoplastic related reconstruction. “Primary tumor site” was classified according to the American Joint Committee on Cancer (AJCC) staging guidelines.¹⁶ “Free flap donor site” was recorded as the site the free flap was harvested from. “Doppler vessel(s)” was recorded as the anastomotic vessel(s) receiving a Doppler probe(s). “Doppler duration” was recorded as the number of postoperative days the probe remained around the vessel for monitoring.

“Doppler status” was documented as either “present” or “changed/absent.” “Suspected flap compromise” was simply recorded as a yes/no variable depending on whether there was suspicion of flap compromise. “Time to exploration” was recorded as the time from the initial suspicion of flap compromise to the time of operative flap exploration. “Exploration findings” documented the intraoperative findings on flap exploration. “Flap salvage” and “flap failure” were simply recorded as yes or no.

Descriptive statistics were used for all variables. Statistical analysis was performed by a biostatistician with a commercially available statistical software package SPSS (SPSS Inc, Chicago, Illinois). Analysis was performed on the entire data set to determine any significant correlations between patient variables and outcome measures. P < .05 was considered statistically significant.

The Cook-Swartz Doppler Flow Monitoring System consists of an implantable, removable, 20-MHz ultrasonic probe with suturable silicon cuff that is used to secure the probe around the adventitia of the venous or arterial pedicle (Figure 1 and Figure 2). The probe is attached to a wire that exits the surgical site through the incision, where it becomes an external wire sutured to the skin by silicon tabs placed around the wire. The external wire is attached through an extension cable to a portable monitor that provides audible real-time monitoring of flap blood flow. The wire and probe are designed to detach from the silicon cuff with minimal tension, once postoperative vascular monitoring is complete.

A consecutive series of 351 patients (244 men and 107 women; mean age, 58.63 years) treated during December 2001 to April 2006 were included. There were 369 separate free flap surgical procedures performed on these patients; 12 patients had repeated surgery for recurrence, 4 had repeated surgery for flap failure, and 1 patient returned twice for non–cancer-related reconstruction. Furthermore, 15 of these patients underwent reconstruction with 2 free flaps, for a total of 384 free flaps monitored using the implantable Doppler probe (Table 1).

The most common indication for surgery was squamous cell carcinoma (81.0%) (Figure 3). Of these, 77 patients underwent surgery for recurrent malignant tumors: 65 had initially chosen nonsurgical treatment and 12 had previous surgical management of their malignant tumor. The distribution of sites reconstructed was as follows: oral cavity (39.8%), oropharynx (30.6%), hypopharynx (8.1%), larynx (7.9%), scalp and facial skin (7.9%), and other (5.7%). The most common free flap used was radial forearm free flap (68.0%) (Figure 4).

The majority of flaps were monitored via Doppler probe placement on the arterial pedicle, with the facial artery (n = 252) and the superior thyroid artery (n = 66) being the most commonly used arteries for anastomosis. Venous Doppler probe placement was used in combination with an arterial Doppler probe placement in 74 patients. Only 4 patients were monitored by venous Doppler probe alone. The facial vein (n = 37) and the external jugular vein (n = 21) were the most commonly used veins for anastomosis. Most patients were monitored with a single Doppler probe (77.8%); however, toward the end of our study, we began to use Doppler probes on both the arterial and venous anastomoses (74 patients). In those patients who received 2 free flaps, we placed a third Doppler probe on their primary free flap's venous anastomosis (5 patients) (Table 2). There was no statistically significant difference in outcomes associated with use of the single- or double-vessel monitoring. On average, the Doppler probe was used for 8.95 days before removal, with a range of 0 to 40 days.

Of the 369 separate flap procedures performed, 40 cases returned to the operating room for exploration for flap compromise. Six of these cases returned to the operating room for a second flap exploration, with a total of 46 explorations for suspected flap compromise. The mean time to operative exploration was 5.43 hours from the initial suspicion of flap compromise (either because of findings from clinical examination or changes in Doppler signal), with a range of 0.0 hours (intraoperative reexploration) to 27.50 hours. Of the 46 operative explorations, 38 were cases of actual compromise, 7 of which were not salvaged. Three of these flaps failed due to vessel thrombosis, 2 due to necrotizing fasciitis, and 2 due to pharyngeal anastomotic breakdown. Our overall flap success rate was 98.1%, with a salvage rate for compromised flaps of 81.6%.

On operative exploration, the most common findings were hematoma formation in 16 flaps (42.1%), vessel thrombosis in 13 flaps (34.2%), and a kinked vessel with flow impediment in 3 flaps (7.9%). All cases of hematoma and kinked vessel were salvaged, and 10 of 13 cases of vessel thrombosis were salvaged (76.9%). Seven flaps were not salvaged; 6 of these failed flaps were fibular free flaps and 1 was an anterolateral thigh flap. A statistically significant correlation between the source of the compromised donor flap and its salvage existed: compromised fibular flaps were significantly less likely to be salvaged than compromised radial or anterolateral thigh flaps (Pearson χ² test, P = .02).

Of the 46 operative explorations, 31 were preceded by an absent or reduced Doppler signal (67.4%). On exploration, 25 of these cases were confirmed as true flap compromise and 6 were false-positive Doppler signals. Five of these false-positive Doppler signals were from probes placed on the venous pedicle. There was a statistically significant correlation between the Doppler signal and the flap's success: flaps that had a change or loss of Doppler signal were significantly more likely to be compromised (Pearson χ² test, P < .001) and ultimately fail (Pearson χ² test, P = .007). Of the cases that were proven to be in true compromise, the mean time to operative exploration was 4.99 hours, with a flap salvage rate of 92.0%.

The remaining 15 cases of flap exploration were not associated with a change in Doppler signal (32.6%). On exploration, 13 of these cases were confirmed as true flap compromise and 2 cases were not (true negatives). However, none of these 13 cases experienced vascular compromise of the monitored vessel: 9 had patent and flowing vascular pedicles, and 4 experienced venous obstruction that was not evidenced by a change in the arterial Doppler signal. For the 13 cases that were proven to be truly compromised, the mean time to operative exploration was 7.67 hours, with a flap salvage rate of 61.5% (Table 3).

In 1998, a study by Kind et al¹ evaluated the use of the implantable Doppler ultrasonic probe in 147 consecutive flap procedures (19 for head and neck reconstruction). The authors reported 20 instances of thrombosis or spasm, with a 100% salvage rate of ischemic flaps. They reported 4 false-positive and no false-negative results.^1,14 A 2003 study by de la Torre et al³ retrospectively evaluated the use of the implantable Doppler probe in 118 patients, 53 of which were cases of head and neck reconstruction. The authors reported 8 instances of anastomotic compromise, with an 83% salvage rate of ischemic flaps. They reported 6 false-positive and no false-negative results.³

A recent retrospective study by Pryor et al¹³ evaluated the use of the implantable Doppler probe in 24 head and neck free flap reconstructions. They reported 3 instances of vessel compromise, with a 66.7% salvage rate. They reported 3 false-positive and no false-negative results for Doppler signaling. The authors also reported using the probe on an artery in half of their cases and found no affect on the quality of monitoring in their patients. However, they suggested the arterial signal was predictably stronger and easier to detect.¹³

While the Cook-Swartz implantable Doppler probe was initially introduced as a venous monitor, our early experience with using the probe on the venous pedicle was unreliable. Of our 7 false-positive results, 5 were due to a lost venous signal in an otherwise healthy flap and thus resulted in a high rate of unnecessary flap exploration. We found that using the probe on the arterial pedicle provided more reliable postoperative monitoring. Previous studies have listed false-positive results as being one of the weaknesses of the implantable Doppler probe as a postoperative monitoring tool.^1,3,15 However, these same studies mostly investigated the implantable probe's use as a venous monitor rather than an arterial monitor. Our experience suggests that arterial use of the monitor results in a lower incidence of false-positive results and therefore an increased specificity for the detection of vascular compromise.

In the cases of flap compromise signaled by a change or loss of Doppler signal (n = 25), the mean time to operative exploration was only 4.99 hours, with a flap salvage rate of 92.0%. For the 13 cases of flap compromise not associated with a change in Doppler signal, the mean time to operative exploration was 7.67 hours, with a flap salvage rate of only 61.5%. However, on operative exploration it was found that none of these cases experienced vascular compromise of the monitored vessel: 9 had patent and flowing vascular pedicles and 4 experienced venous obstruction that was not evidenced by a change in the arterial Doppler signal.

Therefore, use of the Cook-Swartz Doppler Flow Monitoring System resulted in earlier identification of flap compromise and thus earlier operative exploration. This resulted in a 30% increase in flap salvage rates over those flaps that did not experience compromise in the monitored vessel and therefore were not identified by a change in Doppler signal. With operative exploration used as the gold standard, the implantable Doppler probe had a sensitivity of 65.8% and specificity of 98.2%, a positive predictive value of 80.7%, and a negative predictive value of 96.2% in the detection of flap compromise. For the detection of vascular compromise of the monitored vessel (excluding cases involving venous obstruction with an intact arterial Doppler probe, hematoma formation and flap congestion without directly affecting flow in the vascular pedicle, and necrotizing fasciitis without change in vascular flow), the sensitivity increased to 100% and the negative predictive value increased to 100% (Table 4).

Free flap failure is a devastating complication, often requiring salvage surgery and a prolonged hospital stay. Early identification and intervention is well documented to improve the chance of flap salvage.^5,17 Our use of the Cook-Swart Doppler probe allowed us to recognize vascular compromise earlier, resulting in a 92.0% rate of successful salvage in patients with vascular compromise of the monitored vessel. To our knowledge, this is the largest reported series of implantable Doppler probe use in microvascular reconstruction. There were no complications as a result of the implantation or removal of the Doppler probe. A review of our experience would suggest that this is a reliable technique for postoperative monitoring in head and neck reconstruction. A further benefit that we have noted is the intraoperative detection of flap compromise. Often, Doppler signal changes may identify vascular compromise before wound closure and appropriate measures can be taken immediately (repositioning or reanastomosis), sparing the patient a future trip to the operating room.

Our data indicate that although vascular compromise plays a key role in flap failure, it is not the sole cause. Of our 7 flap failures, 5 were from nonanastomotic causes, which indicates the importance of clinical monitoring of the patient. Because our early experience with using the implantable Doppler probe on the venous pedicle was unreliable, currently we do not recommend routine venous monitoring because we believe it may lead to an unacceptably high rate of unnecessary flap exploration. We are presently evaluating this system for simultaneous arterial and venous monitoring in a prospective fashion to provide further guidance on this issue.

In conclusion, the Cook-Swartz Doppler Flow Monitoring System is a valuable adjunct to traditional flap monitoring routines. The implantable Doppler probe is a particularly valuable tool in postoperative monitoring of buried flaps that are not amenable to clinical examination. Furthermore, the audible signal produced by this monitoring system is easy to interpret for support staff and less dependant on clinical experience and judgment than more traditional flap monitoring routines. It can successfully identify vascular compromise at an early stage, allowing high rates of successful flap salvage. We recommend the routine use of the implantable Doppler probe in free flap reconstruction of the head and neck.

Correspondence: Jeffrey R. Harris, MD, FRCSC, Division of Otolaryngology, Head and Neck Surgery, 1E4.29 Walter MacKenzie Health Science Centre, 8440 112th St, Edmonton, AB T6G 2R7, Canada (jeffreyharris@cha.ab.ca).

Submitted for Publication: April 22, 2007; final revision received November 10, 2007; accepted November 13, 2007.

Author Contributions: Dr Harris had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Guillemaud, Cote, and Harris. Acquisition of data: Guillemaud, Seikaly, Cote, Allen, and Harris. Analysis and interpretation of data: Guillemaud, Cote, and Harris. Drafting of the manuscript: Guillemaud, Cote, and Harris. Critical revision of the manuscript for important intellectual content: Guillemaud, Seikaly, Cote, Allen, and Harris. Statistical analysis: Cote and Harris. Obtained funding: Seikaly and Harris. Administrative, technical, and material support: Guillemaud, Cote, Allen, and Harris. Study supervision: Seikaly and Harris.

Financial Disclosure: None reported.

Funding/Support: Support for this project, used for associated administrative costs, paperwork, medical record pulling, and conference costs, was provided by The Head and Neck Oncology Program Fund from the Division of Otolaryngology, Head and Neck Surgery, University of Alberta, Edmonton, Alberta, Canada.

Previous Presentation: This study was presented at the American Head and Neck Society 2006 Annual Meeting; August 18, 2006; Chicago, Illinois.

Additional Contributions: William K. Midodzi, MSc (biostatistics), PhD (epidemiology), Epidemiology Coordinating and Research (EPICORE) Centre, The University of Alberta Hospital, Edmonton, Alberta, Canada, contributed to this study as a biostatistician.