This article discusses the lessons learned from nearly 2700 free tissue transfer procedures to reconstruct defects of the head and neck at the University of Washington. It discusses the authors’ perioperative management practices regarding perioperative tracheotomy tube placement, their method of postoperative flap monitoring, and their current protocol for use of postoperative antibiotics. It reports on the reconstructive preferences for 2 difficult defects that frequently result in unfavorable outcomes: the total glossectomy defect and the pharyngolaryngectomy defect. Key points for harvesting and insetting flaps, to maximize reconstructive outcomes, are provided.
Key points
- •
Overnight intubation has allowed a decrease in the use of tracheotomy tube placement without adverse patient outcomes. The authors have found earlier return to normal swallow, decreased intensive care unit stay duration, and decreased overall hospitalization.
- •
Current practice is to follow the Infectious Disease Society of America guidelines for perioperative antibiotic use with ampicillin-sulbactam or clindamycin with levofloxacin (or another agent for broad-spectrum gram-negative coverage) for 24 hours postoperatively.
- •
For postoperative flap monitoring, the authors use a needle-stick technique every 6 hours for 72 hours postoperatively. For buried flaps without a monitor paddle, the authors use both arterial and venous Cook implantable Dopplers.
- •
For total glossectomy defects, our preference has evolved to use the anterolateral thigh flap for its bulk, minimal atrophy, and large surface area. We use the fascia of the vastus lateralis to suspend the flap to the mandible and perform hyoid suspension concomitantly.
- •
For total cervical esophagectomy defects, our preference is to use jejunal free flap secondary to improved swallow outcomes with decreased fistula and stricture rates. We have found the voice changes to be inconsequential to patients.
Introduction
Free tissue transfer to the head and neck is a complex and multistep procedure. With experience, clinicians can become proficient with the reconstructive technique, but achieving consistently reliable functional and cosmetic results is a challenge even for the most seasoned surgeon, particularly when reconstructing head and neck oncologic defects, which are frequently seen in previously irradiated patients with complex functional needs and/or multiple medical comorbid conditions.
At the University of Washington (UW), the senior author has performed more than 2500 free tissue transfers to the head and neck over the past 20 years. This article shares our experience with 3 areas that we find particularly challenging in the management of our patients. The article describes our approach to postoperative care, because most unfavorable outcomes arise in the early postoperative period. It then describes 2 complicated defects: the total glossectomy defect and the total laryngopharyngectomy defect. These 2 defects result in severe form and functional losses for patients and are challenging to adequately restore.
Introduction
Free tissue transfer to the head and neck is a complex and multistep procedure. With experience, clinicians can become proficient with the reconstructive technique, but achieving consistently reliable functional and cosmetic results is a challenge even for the most seasoned surgeon, particularly when reconstructing head and neck oncologic defects, which are frequently seen in previously irradiated patients with complex functional needs and/or multiple medical comorbid conditions.
At the University of Washington (UW), the senior author has performed more than 2500 free tissue transfers to the head and neck over the past 20 years. This article shares our experience with 3 areas that we find particularly challenging in the management of our patients. The article describes our approach to postoperative care, because most unfavorable outcomes arise in the early postoperative period. It then describes 2 complicated defects: the total glossectomy defect and the total laryngopharyngectomy defect. These 2 defects result in severe form and functional losses for patients and are challenging to adequately restore.
Postoperative management: minimizing the morbidity of head and neck free flap surgery
Convalescence from head and neck free flap surgery is a challenging endeavor. Recent studies suggest that between 11% and 52% of patients with head and neck cancer meet Diagnostic and Statistical Manual of Mental Disorders criteria for major depressive disorder during treatment and recovery. As a result, some recent investigators have suggested the prophylactic initiation of antidepressant therapy for all patients with head and neck cancer because of the common and significant physiologic insults caused by either surgery or postoperative chemoradiotherapy. In the postoperative period, significant facial/neck swelling, restriction of oral intake, loss of normal vocalization, frequent blood draws, and invasive procedures contribute to a high rate of discomfort and, at times, hopelessness. In addition to the development of a free flap clinical care pathway, the authors have focused on 3 aspects of postoperative management that, in our experience, have minimized morbidity and improved the quality of care: reduction in tracheostomy tube placement, prevention of postoperative infection while minimizing side effects and complications associated with antibiotic use, and reduction in the risk of free flap loss with dependable monitoring.
Development of a Clinical Care Pathway
The perioperative care of all free flap patients at UW is guided by a streamlined clinical care pathway. This pathway begins preoperatively with significant counseling and an assessment of anticipated postoperative social work/nursing needs. Our head and neck cancer care team consents patient before surgery for peripherally inserted central catheter (PICC) and percutaneous endoscopic gastrostomy (PEG) placement; procedures that are performed while the patient is still anesthetized in the intensive care unit (ICU) the morning after surgery. PEG placement ensures that patients receive enteral tube feeds and medications within 24 hours of surgery, allowing the decreased use of intravenous fluids and avoiding the discomfort of a nasal tube. The PICC allows most new needle sticks to be avoided during their stay. Although these measures do not eliminate the recovery burden, anecdotally, they make the hospitalization and immediate home care much more tolerable. In addition, early ambulation and care by a focused head and neck nursing team has resulted in a 2-day reduction of hospital stay (average now 8.2 days) and increased levels of patient satisfaction within the first 6 months of implementation of the care pathway.
Overnight Intubation Versus Tracheostomy
Tracheostomy has traditionally been considered the mainstay of airway management following free tissue transfer to the head and neck. Many patients require this type of secure airway to avoid life-threatening obstruction; however, tracheostomy tube placement involves morbidity and complications. Therefore, at UW, we have shifted the airway management paradigm to reduce tracheostomy-related morbidity by avoiding tracheostomy in most of our head and neck free flap patients.
Tracheostomy tube complications include bleeding, infection, mucous plugging, tracheitis, aspiration-related pneumonia, and tracheal stenosis. Data suggest that there may be an increased risk of pulmonary complications in head and neck patients with tracheostomies, because tracheostomy is known to exacerbate aspiration, although it does improve pulmonary toilet. Avoiding a tracheostomy allows patients to speak earlier and maintain a strong cough. A tracheostomy may require increased nursing care and patient education and frequently complicates patients’ placement after surgery, because many US skilled nursing facilities are disinclined to approve the transfer of patients with tracheostomy tubes.
In a 2010 retrospective review published by our UW group, 37 patients who were nasally intubated following an oral cavity resection/free flap reconstructions were compared with 21 patients who underwent a tracheostomy following similar oral cavity reconstruction. The mean hospital stay (8.4 days vs 12.4 days) and the likelihood of requiring a feeding tube (19% vs 76%) at discharge were both independently increased in the tracheostomy group on multivariate analysis. There were no airway emergencies or secondary tracheostomies performed in the nasal intubation group.
Similarly, there seems to be a growing trend elsewhere in the literature toward the decreased use of tracheostomy tubes in patients undergoing head and neck free tissue transfer. In a 2014 report by Moubayed and colleagues, 66 patients underwent mandibulectomy with osteocutaneous free flap reconstruction without tracheostomy. There were no deaths or major safety events in the series, although 1 patient required an urgent secondary tracheostomy and 2 patients reported airway obstructive symptoms. A 2013 study by Coyle and colleagues reported outcomes of 50 patients undergoing free flap reconstruction with oncologic resections without a tracheostomy and a similar group of 50 patients who underwent an elective tracheostomy. There were no airway emergencies or secondary tracheostomies reported in the intubation group. In addition, the intubation group had a shorter ICU stay (1.4 days vs 3.7 days), shorter overall hospital stay (13 days vs 18 days), and a lower rate of pneumonia (10% vs 38%).
Since our 2010 publication, it has become our preference to avoid a tracheostomy in most patients undergoing free flap reconstruction. Defects for which we prefer overnight intubation include most defects of the oral cavity: large reconstructions of the oral tongue, and lateral and anterior floor of mouth; segmental mandibulectomy; as well as maxillectomy, external skin, skull base, or parotidectomy defects. Patients who we think will still benefit from an elective tracheostomy include patients who require free tissue transfer for total glossectomy defects, oropharyngeal defects (both pharyngeal and base of tongue), and hypopharyngeal defects. We are also prone to consider a tracheostomy in patients with very poor lung function, trismus, and bulky flaps that would significantly complicate reintubation. If the ease of reintubation is in question, we frequently assess laryngeal exposure with a postoperative direct laryngoscopy before the patient leaves the operating room.
Patients for whom a tracheostomy is deferred receive intravenous steroids (Decadron 10 mg intravenously every 8 hours for 24 hours) while in the ICU overnight, and then they are extubated on the first postoperative morning, unless significant airway swelling has developed. Using this approach, only a very few patients have required further airway management.
Postoperative Infections
One important clinical challenge following head and neck free tissue transfer is the high rate of postoperative infections. The location and nature of head and neck cancer resection often directly violates oral/pharyngeal surfaces and thus exposes wounds to a significant bacterial burden. This exposure places patients at high risk of wound infection, with an incidence of 8% to 45%. Limiting the rate of infection is critical in free flap surgery, because infection increases the risk of wound complications and microvascular failure. However, a consensus on the type and duration of antibiotic prophylaxis has not been uniformly established. In a landmark 1984 study, Johnson and colleagues reported that the rate of infection following head and neck cancer surgery, which was highest in oropharyngeal resections and free flap reconstructions, was greatly reduced by treating patients with 5 days of postoperative clindamycin and gentamicin (broad-spectrum gram-positive/gram-negative coverage). In 2014, the same author performed a literature review and proposed evidence-based recommendations for treatment with 24 hours of perioperative antibiotic therapy for clean-contaminated head and neck surgery. This proposal is consistent with current guidelines from the Infectious Disease Society of America (IDSA), Surgical Infection Society, and Society for Healthcare Epidemiology of America, all of which recommend 24 hours of antibiotic prophylaxis for major clean-contaminated procedures.
The authors recently analyzed and published changes to our prophylactic antibiotic protocol for patients undergoing head and neck free tissue transfer procedures. As we transitioned from prolonged courses of antibiotics, as Johnson and colleagues initially published, to the IDSA-recommended 24-hour duration, our data showed that within 30 days of surgery (which is the US Centers for Disease Control and Prevention defined time period for surgical site infection) 45% of the patients were treated for at least 1 infection: 22% at the flap inset site, 12% at donor site, and 17% at any other site (eg, pneumonia, urinary tract infection). The rate of infection in patients who received 24 hours of antibiotics was higher at 57% compared with a 42% rate in patients who received prolonged courses of antibiotics (most commonly 7 days) after surgery ( P = .01). Use of clindamycin alone relative to ampicillin-sulbactam increased the risk of postoperative infection (odds ratio [OR], 2.5; P = .01), which is presumably from limiting gram-negative bacterial coverage in patients who received clindamycin alone. In multivariate analysis, the other factor most strongly associated with postoperative infection was hypothyroidism (OR, 10.8; P <.001).
Some of these findings are in conflict with other recent studies, including a 2015 study by Khariwala and colleagues showing an increased rate of pneumonia in patients receiving longer courses of antibiotics with no reduction in surgical site infection in patients undergoing radial forearm flap reconstruction following a laryngectomy. In their analysis, obesity was the greatest risk factor and had the highest OR for developing a postoperative infection.
Postoperative antibiotic use does involve harm. The increasing frequency of antibiotic-associated complications (eg, drug reactions, Clostridium difficile infections), as well as the emergence of increasing multidrug-resistant bacteria, are intuitive reasons to avoid prolonged courses of antimicrobial therapy. However, recent data on the duration of perioperative antibiotic use are sparse in the head and neck free flap population. In our study, there was no clear correlation between antibiotic duration and C difficile infection or multidrug-resistant infection. However, this may be institutionally dependent.
In conclusion, data from our institution show a high rate of infection following head and neck free flap surgery despite adherence to the IDSA perioperative guidelines. Contributing to this high rate may be our inclusion of the full 30-day observation period, or it may represent our bias toward having a very low threshold for diagnosing postoperative infection and initiating treatment. Despite this result, our current preference is to adhere to the shorter IDSA-recommended 24 hours of broad-spectrum antimicrobial prophylaxis using ampicillin-sulbactam (first line) or clindamycin/levofloxacin (for gram-negative bacteria coverage) in patients with ampicillin allergy. The authors continue to maintain a very low threshold for prolonged antibiotic use (5 to 7 days) for patients with hypothyroidism or poor wound healing given supportive data from our recent report.
Postoperative Free Flap Monitoring
More than 20 different methods for free flap monitoring have been described in the literature. Methods range from a simple clinical examination and pinprick testing to using complex technological innovations, including methods using implantable Doppler ultrasonography, microdialysis probes, oxygen saturation probes, laser Doppler, fluorometry, implantable temperature probes, glucose/lactate monitors, handheld Doppler, scintigraphy, pH monitoring, impedance plethysmography, confocal microscopy, smartphone serial photoimaging, and several other techniques.
Despite this variety, the goals of monitoring remain singularly focused on reducing the rate of free flap loss. Monitoring techniques achieve this goal by heralding the risk of free flap compromise before overt clinical change, thus improving the chances of free flap salvage. Timely identification of a threatened flap increases the rate of salvage and decreases the risk of the no-reflow phenomenon.
Despite a high rate of free flap survival, clinicians at UW continue to place a high priority on meticulous monitoring to ensure our high rate of success. Over the past 20 years, many methods for free flap monitoring have been tried by the authors, but in our practice no method has been as reliable and accurate as simple skin prick testing. Our protocol calls for a nursing-performed clinical examination every hour for the first 24 hours while patients are in the ICU overnight. In addition, the flap is checked by a physician with a clinical examination and pinprick testing every 6 hours for a total of 72 hours. After 72 hours, the flap is checked twice daily without planned pinprick until discharge.
In order to facilitate pinprick monitoring, a concerted effort is made to develop a monitoring paddle during reconstruction. If a monitoring paddle is not used (a completely buried flap), then we use an implantable Doppler system on both the artery and the venous outflow. This use of the Cook implantable Doppler on buried flaps has been shown to be particularly valuable, as was shown in a 2011 retrospective review of 548 microsurgical cases in which the rate of salvage for buried flaps was improved significantly with use of the implantable Doppler versus a clinical examination alone (94% vs 40%). Continuous implantable Dopplers are also preferred following free flap take-backs that require revision of the microvascular anastomosis.
Fig. 1 shows the use of pinprick testing with a 25-gauge needle. A reassuring result is shown by the delayed (2 to 5 seconds) appearance of bright red blood. Fig. 2 shows a flap that is congested on pinprick testing. The 46-year-old patient underwent a composite oral cavity resection followed by reconstruction with an osteocutaneous radial forearm flap. On the first postoperative day, the patient was noted as having dark venous blood on pinprick testing. A subsequent take-back was significant for the twisting of the venous outflow tract causing venous obstruction. A revision venous anastomosis was performed and no further issues developed, resulting in successful free flap salvage.
Using the pinprick method with judicious use of the implantable Doppler has generally yielded a high rate of free flap success and salvage. A review of an unpublished personal database by the senior author (NDF) reveals an overall flap success rate of 99.2% in 2734 free flaps. The corresponding take-back rate is 4.1% with a salvage rate of 91%. There is a partial flap necrosis rate of 6.6%.
Lesson learned from reconstruction of the total glossectomy defect
The total glossectomy defect (with the preservation of the larynx) represents one of our most significant reconstructive challenges. Loss of tongue function, even with adequate tissue restoration, is a serious consequence for patients, resulting in a propensity for aspiration as well as distortion of articulation and swallowing. A total glossectomy is generally performed in patients who are severely malnourished and often previously irradiated (oropharyngeal). Loss of flap bulk and the lack of long-term suspension of the neotongue frequently diminish long-term reconstructive results. This loss of neotongue volume has repeatedly been shown to result in poorer swallowing and voice outcomes.
The appropriateness of a total glossectomy with laryngeal preservation was controversial until the 1980s to 1990s. Before this, controversy existed as to whether head and neck surgeons should perform concurrent total laryngectomy in all patients undergoing a total glossectomy because of concerns of poor residual laryngeal function. Before the widespread advent of free tissue transfer, total glossectomy defects were generally reconstructed with the pectoralis flap. This flap had a significant issue with maintaining bulk in the oral cavity and frequently receded into the neck with gravity during recovery and atrophy. Swallowing results were poor and many patients were left with an essentially nonfunctional larynx (if laryngeal preservation was attempted). The evolution toward improved free tissue reconstructive techniques has decreased the need for total laryngectomy. Despite these reconstruction improvements, patients remain at a high risk for severe dysphagia, aspiration, and poor articulation.
In theory, the ideal flap choice for reconstructing the total glossectomy defect would be designed with significant oversized bulk, would not atrophy significantly over time, and would remain suspended in the oral cavity. Flap bulk is critical to the oral phase of swallowing. In order to maintain a functional oral swallow, the neotongue must remain in contact with the palate to allow the patient to maintain an adequate suction with oral swallowing. This palatal contact requires that the flap not undergo a significant loss of bulk and maintains adequate suspension. In addition, it is necessary to have significant oropharyngeal/base of tongue bulk to divert the food/liquid bolus away from the glottis during the oropharyngeal phase of swallowing, rather than funneling the bolus toward the laryngeal inlet, which is frequently associated with undersized flaps.
In our experience, the radial forearm and lateral arm free flaps do not maintain enough bulk to achieve the volume needed for the neotongue. Fig. 3 shows lack of bulk following lateral arm free flap reconstruction of a total oral tongue defect. The rectus abdominis and latissimus dorsi free flaps are more appropriate choices, capable of the necessary volume recreation. However, we have found that, given their significant muscle bulk, they tend to atrophy significantly with time. These flaps are secondary choices in our practice. Our first choice of a free flap donor site is the anterolateral thigh (in patients with suitable subcutaneous tissue of the upper leg). This flap is typically ideal because it has significant adipose tissue, a strong fascial layer from the vastus lateralis for suspension, and a muscular cuff that can be harvested to provide additional bulk.
