Agra IMG, Carvalho AL, Pontes E, Campos OD, Ulbrich FS, Magrin J, Kowalski LP. Postoperative Complications After En Bloc Salvage Surgery for Head and Neck Cancer. Arch Otolaryngol Head Neck Surg. 2003;129(12):1317-1321. doi:10.1001/archotol.129.12.1317
Copyright 2003 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2003
To analyze the frequency of and risk factors for postoperative complications after en bloc salvage surgery for head and neck cancer.
Retrospective cohort study.
Patients were evaluated from February 7, 1990, to November 17, 1999, in a tertiary cancer center hospital.
Consecutive sample of 124 patients from the hospital database. Only patients with recurrent head and neck squamous cell carcinoma undergoing en bloc salvage resection were eligible for the study.
Main Outcome Measures
We analyzed the frequency of and risk factors for complications after salvage surgery.
The tumor location was the lip in 6 patients, oral cavity in 55, oropharynx in 31, larynx in 24, and hypopharynx in 8. Previous treatment was surgery alone in 20 patients, radiotherapy alone in 68, surgery and radiotherapy in 21, and radiotherapy and chemotherapy in 14. An additional patient received chemotherapy alone before salvage surgery. The clinical stage of the recurrent tumor was I or II in 23 patients and III or IV in 101 patients. Postoperative complications occurred in 66 patients (53.2%). Fifty-three patients (42.7%) had minor complications, and 23 patients (18.5%) had major ones. There were 4 postoperative deaths (3.2%). The major factor associated with the overall occurrence of postoperative complications was the clinical stage of the recurrent tumor (P = .02). The occurrence of minor complications correlated with the previously treated site, with complications occurring more often in patients undergoing locoregional vs local treatment (P = .04). Major complications were associated with the time between initial treatment and salvage surgery (P = .05).
Salvage surgery can be performed with acceptable rates of postoperative complications. The clinical stage of the recurrent tumor and the previous site treated were the 2 major factors associated with the occurrence of postoperative complications.
SQUAMOUS CELL carcinomas are the most frequent neoplasms of the upper aerodigestive tract, and the diagnosis is generally made at advanced clinical stages, with a high risk of locoregional recurrence after treatment.1- 4 The poor prognosis of these recurrent cancers, the high risk of postoperative complications, and the high cost of the procedure render salvage treatment controversial.5 Moreover, the indications for this treatment vary widely in the literature. Some authors recommend less extensive salvage procedures,6,7 while others advocate major surgical procedures with complex reconstructions.5,8- 10
Recent studies11,12 have assessed the efficacy of chemotherapy and radiotherapy in some patients with advanced disease, reserving salvage surgery for patients who do not respond to treatment or who have recurrences. This approach is based on a theory of organ preservation in patients with squamous cell carcinoma of the larynx and hypopharynx and in selected patients with oropharynx tumors. These studies found overall survival rates to be similar to those obtained in patients who undergo surgery initially. However, increased complications are associated with salvage surgery compared with initial primary surgery.9,13,14
The objective of this study was to evaluate the factors associated with the occurrence of postoperative complications in head and neck cancer patients undergoing en bloc salvage surgery.
From February 7, 1990, to November 17, 1999, salvage surgical procedures were performed in 324 patients for treatment of recurrent squamous cell carcinomas of the upper aerodigestive tract at the Hospital do Câncer A. C. Camargo. The inclusion criteria for this study were histological confirmation of recurrent squamous cell carcinoma, exclusion of the possibility of a second primary tumor and distant metastasis, and en bloc surgery with curative intent (resection of the primary tumor and unilateral or bilateral neck dissection).
One hundred twenty-four patients were eligible; 105 (84.7%) were males and 19 (15.3%) were females. Ages ranged from 15 to 83 years (median, 56 years). The primary tumor sites were the lip in 6 patients, oral cavity in 55, oropharynx in 31, larynx in 24, and hypopharynx in 8. Previous treatment consisted of surgery in 20 patients (16.3%), surgery followed by radiotherapy in 21 patients (17.1%), radiotherapy alone in 68 patients (55.3%), and radiotherapy and chemotherapy in 14 patients (11.4%). One patient received chemotherapy alone before the salvage surgery. We grouped patients according to the site of previous treatment: 30 patients (24.4%) had initial treatment to the primary tumor only (local treatment), and 93 (75.6%) had treatment to the primary tumor and the neck (locoregional treatment).
Recurrences were staged according to the TNM classification.15 Twenty-three patients had clinical stage I or II recurrences, and 101 had stage III or IV recurrences. The interval between initial treatment and salvage surgery ranged from 2.4 to 267.8 months (median, 10.6 months).
Antibiotic prophylaxis with clindamycin phosphate plus amikacin sulfate is routinely given at anesthesia induction and for the subsequent 24 hours. The study group underwent 179 surgical procedures: 22 partial glossectomies (17.7%), 48 glossomandibulectomies (38.7%), 7 buccopharyngectomies (5.6%), 55 commando procedures (44.4%), 10 wide resections of carcinomas of the oropharynx via paramedian mandibulectomies (8.1%), 5 horizontal supraglottic laryngectomies (4.0%), 28 total laryngectomies or pharyngolaryngectomies (22.6%), and 4 total glossolaryngectomies (3.2%). Neck dissections were bilateral in 46 cases. Homolateral radical neck dissection was performed in 75 cases (5 extended) and contralateral dissection in 13 cases. Homolateral selective neck dissection (supraomohyoid or lateral) was performed in 39 cases and contralateral dissection in 43 cases. Among the study group, the type of reconstruction used was primary closure in 45 patients (36.3%), local flaps in 17 (13.7%), myocutaneous flaps in 58 (46.8%) (51 [41.1%] consisting of pectoralis major), and free flaps in 4 (3.2%).
The outcomes were divided into minor and major complications. Minor complications were non–life-threatening local complications without need for reinterventions. Major complications included life-threatening systemic and local complications, such as rupture of great vessels, chyle fistulas, hematomas, and large fistulas requiring surgical reinterventions.
Statistical analysis was performed on a database using SPSS for Windows (version 10.0; SPSS Inc, Chicago, Ill). Postoperative death was defined as death directly related to and occurring within the first 30 days after surgery. The mean lengths of hospital stay were compared by t test. Associations between the clinical, pathological, and therapeutic variables and the risk of complications were assessed by χ2 test or Fisher exact test when applicable. Logistic regression was calculated to evaluate the independent risk factors for postoperative complications. Statistical significance was set at P≤.05.
The overall rate of postoperative complications was 53.2% (66 of 124 patients). Complications were minor in 53 patients (42.7%) and major in 23 (18.5%). Thirty patients (24.2%) had more than 1 complication, most commonly fistulas, infections, and wound dehiscence. Ten patients (8.1%) had both minor and major complications.
The length of hospital stay ranged from 1 to 90 days (mean, 11.2 days). The mean length of stay was 7.4 days for patients without surgical complications, compared with 14.5 days for those with complications (P<.01) (Table 1). The length of stay also varied based on whether the complications were minor (mean, 12.3 days) or major (mean, 18.7 days) (P = .04).
Seventy-six patients (61.3%) received a temporary tracheotomy, and 32 (25.8%) had a total laryngectomy. The duration of the temporary tracheotomy ranged from 4 to 500 days (median, 48.5 days). In 16 (21.1%) of the 76 patients, the temporary tracheotomies were constructed with no intent of removal. All patients had a nasoenteral feeding tube in place during the postoperative period, ranging from 10 to 600 days (median, 47 days). In 34 patients (27.4%), the tube was still in place at death or at the last follow-up visit.
The most frequent surgical complications were wound infections in 31 patients (25.0%), salivary fistulas in 28 (22.6%), wound dehiscence in 14 (11.3%), and partial flap necrosis in 12 (9.7%). Life-threatening local complications included total necrosis of the myocutaneous flap in 2 patients (1.6%) and rupture of the carotid artery in 2 patients (1.6%). Seven patients (5.6%) were operated on for large salivary fistulas, mainly associated with infections and wound dehiscence (Table 2). In all of these patients, flaps were performed for the surgical repair, primarily pectoralis major flaps (in 6 patients). Management of a wound infection and a minor partial flap necrosis was conservative, with antibiotics used as necessary based on persistence of drainage and culture findings.
Patients with rupture of the carotid artery underwent surgical interventions. One patient had carotid artery ligation, experienced a stroke, and died. In a second patient, ligation of the common carotid artery was performed, a pectoralis major flap was constructed to protect the artery, and the patient recovered without additional sequelae.
Two patients with total flap necrosis also underwent surgical interventions. In one patient, another myocutaneous flap operation was performed, and in the other a free flap was constructed. Both patients recovered without further complications.
Four patients (3.2%) died within the postoperative period. Two died of systemic complications (1 stroke and 1 pneumonia), and another died of a local infection that evolved into sepsis. The other patient died of carotid artery rupture.
Neither age nor sex correlated with the development of minor or major surgical complications. The site of the primary tumor, which indirectly reflects the type of surgery, was also not associated with a risk of postoperative complications. Patients with advanced clinical stage recurrences (stages III and IV) showed higher overall rates of postoperative complications, 58.4% and 30.4%, respectively (P = .02). Among patients who initially underwent local treatment to the primary tumor only, by surgery or radiotherapy, 26.7% had postoperative minor complications, while 48.4% of patients undergoing locoregional therapy to the primary tumor and the neck had postoperative minor complications (P = .04). The type of previous treatment was associated with a risk of postoperative minor complications, with the patients who were previously treated with radiotherapy combined with chemotherapy experiencing a higher, although not statistically significant (P = .08), rate of minor complications (Table 3). The time between initial treatment and salvage surgery was significantly associated with a risk of major complications (P = .05).
On multivariate analysis, an independent risk factor for overall postoperative complications was advanced clinical stage (stages III and IV) of the recurrent tumor (odds ratio, 3.0; 95% confidence interval, 1.1-8.0). For minor postoperative complications, advanced clinical stage was also an independent risk factor (odds ratio, 3.1; 95% confidence interval, 1.0-9.2), as was the previous site treated (local tumor plus the neck) (odds ratio, 2.6; 95% confidence interval, 1.0-6.5). Patients with longer than 1 year between initial treatment and salvage surgery were at increased risk of major complications (odds ratio, 2.5; 95% confidence interval, 1.0-6.3).
We studied patients who underwent salvage surgery by en bloc resection to assess morbidity and mortality outcomes, in what would be considered clean-contaminated surgical procedures. En bloc salvage surgical procedures for the treatment of recurrent squamous cell carcinomas of the upper aerodigestive tract leave saliva in communication with the neck wound, increasing the risk of complications such as fistulas and wound infections. These can lead to other complications such as wound dehiscence, flap necrosis, and life-threatening rupture of the carotid artery. For classification of minor and major complications, we used criteria similar to those described by Sassler et al16 and Bengston et al.10 The only difference was that we also considered systemic complications, not just local complications, as they did.
In our series analyzing patients undergoing en bloc salvage surgical procedure, the overall rate of postoperative complications was 53.2%, and postoperative death occurred in 3.2% of the patients. Our postoperative complication rates were similar to those in other studies.2- 4 In a series of 114 patients with squamous cell carcinoma of the retromolar trigone, which included the commando procedure as a primary treatment, an overall rate of postoperative complications of 51.8% and a postoperative mortality of 0.9% were observed.17 Gehanno et al14 reviewed their use of the commando procedure in 120 patients with squamous cell carcinoma of the tonsil (70 undergoing primary surgical treatment and 50 undergoing salvage surgical procedures after the failure of radiotherapy) and found similar rates of postoperative complications but higher postoperative mortality in the salvage surgery group (1.4% vs 8%). Girod et al18 studied 159 patients, among whom 134 underwent en bloc surgical procedures for squamous cell carcinomas of the upper aerodigestive tract, and found an overall postoperative complication rate of 62.9%, with wound infections and fistulas being the most frequent complications. Postoperative mortality was 1.2%, and radiotherapy before surgery was a significant risk factor for postoperative complications.
In this series, age was not a risk factor for postoperative complications; postoperative morbidity was similar among patients 65 years and older and those younger than 65 (Table 3). In a case-control study comparing morbidity and survival associated with head and neck surgery in patients 70 years and older and those younger than 70, no significant differences were found as well.19 Clayman et al20 published similar results in patients 80 years and older. In our study, the clinical stage of tumor recurrence was associated with a higher rate of postoperative complications, and this finding was also previously reported by other authors.18
In our series, 104 patients (83.8%) had previously received radiotherapy, 14 in combination with chemotherapy. Previous treatment with radiotherapy was associated with a higher risk of postoperative minor complications; 78.6% of patients with combined chemotherapy experienced complications, compared with 47.1% who received radiotherapy alone and 28.6% who underwent surgery followed by adjuvant radiotherapy. Corey et al21 studied the association between chemotherapy and surgical complications in a randomized prospective study in which neoadjuvant methotrexate and leucovorin calcium were administered before surgery. Outcomes of 23 patients who received chemotherapy were compared with those of 19 patients whose primary treatment was surgery. Sixteen of the 23 receiving chemotherapy experienced postoperative complications, compared with 8 of the 19 who did not receive chemotherapy (P<.01). Sassler et al16 reported an overall rate of postoperative complications of 61.0% in patients included in an organ preservation protocol who received induction chemotherapy and radiotherapy. Postoperative deaths occurred in 2 (11.1%) of the 18 patients. Also in their study, surgery performed within the first year of ending chemoradiotherapy was associated with a complication rate of 77.0%, compared with 20.0% among those undergoing salvage surgery after the first year.
Lavertu et al22 published their experience with surgical procedures performed on patients who received radiotherapy alone vs concurrent chemotherapy for the treatment of head and neck carcinomas. The overall rate of postoperative complications was 45.8%. There was no difference between the 2 groups regarding postoperative morbidity and mortality. However only 3 patients had en bloc resection as part of the salvage treatment in this study.
Because patients in whom surgery is indicated after treatment with chemotherapy and radiotherapy are subject to a higher risk of postoperative complications, special care must be taken with these patients. Antibiotic therapy should cover gram-positive and gram-negative bacteria. The dissection of the tissues and hemostasis should be done meticulously, avoiding excessive cauterization and mass ligatures, taking particular care with cutaneous flaps. When needed, well-vascularized flaps should be used for covering the sutures of the pharynx and oral cavity. These flaps, including myocutaneous and free tissue transfers, give extra protection to the carotid artery. Weisman and Robbins13 also suggested pharyngostomas in selected cases, to direct the pharyngeal-cutaneous fistula and to diminish the risk of infection and other local complications, such as flap necrosis and carotid artery rupture. We do not routinely perform pharyngostomas. Our personal preference is a wound closure without tension or the use of a well-vascularized myocutaneous flap or a free tissue transfer. Adequate nutritional support appears to be imperative for a good surgical result, preferably enteral supplementation.19
In conclusion, in our series of patients undergoing en bloc salvage surgery, we observed an acceptable rate of morbidity and mortality. Patients with advanced clinical stage recurrences and those who previously received radiotherapy (especially when combined with chemotherapy) should be carefully managed regarding their salvage surgery, as they are at high risk for postoperative complications.
Corresponding author: Luiz P. Kowalski, MD, PhD, Head and Neck Surgery and Otorhinolaryngology Department, Hospital do Câncer A. C. Camargo, Rua Professor Antônio Prudente 211, 01509-900 São Paulo, Brazil (e-mail: email@example.com).
Submitted for publication April 30, 2002; final revision received November 19, 2002; accepted April 24, 2003.