ORIGINAL ARTICLE. Comorbidity as a Major Risk Factor for Mortality and Complications in Head and Neck Surgery

Transcription

1 ORIGINAL ARTICLE Comorbidity as a Major Risk Factor for Mortality and Complications in Head and Neck Surgery Marciano B. Ferrier, MD; Emiel B. Spuesens; Saskia Le Cessie, PhD; Robert J. Baatenburg de Jong, PhD Objective: To describe the impact of comorbidity on complications of surgery and mortality in patients with head and neck squamous cell carcinoma (HNSCC). Design: A total of 120 consecutive patients with HNSCC, treated surgically between January 1999 and December 2001, were included. The Adult Comorbidity Evaluation 27 index (ACE-27) and the American Society of Anesthesiologists (ASA) risk classification system were used to describe comorbidity. Major complications were defined and scored by review of the medical records. Univariate and multivariate analyses were performed to determine the impact of 17 clinical variables, including the ACE-27 grade and the ASA class. Results: Twenty-five patients (21.4%) had 1 or more major complications. In the univariate analysis, ACE-27 grade, ASA class, T stage, surgical procedure used for the primary tumor, type of neck dissection, and duration of anesthesia had a significant relation with major complications. In the multivariate analysis, duration of anesthesia and comorbidity reflected by the ACE-27 grade or the ASA class remained significant. The odds ratios (95% confidence intervals) associated with ACE-27 grades of 1 and 2 were 1.9 ( ) and 4.6 ( ), respectively; with ASA classes 2 and 3, 2.0 ( ) and 10.0 ( ), respectively. Duration of anesthesia longer than 360 minutes was characterized by an odds ratio of 7.8 ( ). Conclusions: Duration of anesthesia and comorbidity reflected by the ACE-27 grade and the ASA class are important predictors of major complications in head and neck surgery. Optimizing the general condition of patients with HNSCC might reduce morbidity and treatment-related costs. Arch Otolaryngol Head Neck Surg. 2005;131:27-32 Author Affiliations: Departments of Otolaryngology Head and Neck Surgery (Drs Ferrier and Baatenburg de Jong and Mr Spuesens) and Medical Statistics (De Le Cessie), Leiden University Medical Center, Leiden, the Netherlands. Financial Disclosure: None. ONE OF THE TREATMENT OPtions for patients with head and neck squamous cell carcinoma (HNSCC) is surgery. Surgical procedures in the head and neck have a substantial complication rate. Little is known about the risk factors for complications and mortality. Potentially, reliable prediction of complications and mortality could lead to correction of contributing factors. Therefore, evaluation of potential risk factors is of medical and economic importance. This study was undertaken to identify risk factors for complications and mortality. One of the potential risk factors is comorbidity, and a significant number of patients with HNSCC have comorbid disease. 1 Many comorbidity indexes are available, but none is applied universally. The Adult Comorbidity Evaluation 27 index (ACE-27) (Table 1) was developed by Piccirillo 2 and is a modification of the Kaplan-Feinstein Index. 3 The ACE-27 is a validated index especially designed for comorbidity measurements in patients with cancer. A disadvantage of the ACE-27 is that it can be time-consuming. The American Society of Anesthesiologists (ASA) risk classification system (Table 2) is actually an index for perioperative risk, but it can also be used to evaluate comorbidity because it describes a patient s physical status prior to surgery. The ASA class is determined before any patient undergoes anesthesia and is therefore readily available. This study was primarily performed to identify risk factors for complications and mortality in head and neck surgery. Another goal was to compare the comprehensive ACE-27 index with the concise ASA system for predicting complications and mortality. In addition, potential factors influencing the length of hospitalization were studied. METHODS PATIENTS The study population consisted of 120 consecutive surgically treated patients with HNSCC 27

2 Table 1. A Selection of the Adult Comorbidity Evaluation 27 System* Cogent Comorbid Ailment (Cardiovascular System) Grade 3 Grade 2 Grade 1 MI 6 moago 6 mo ago Old MI detected by ECG only; age undetermined Angina/coronary artery disease Unstable angina Chronic exertional angina ECG, stress test, or catheterization evidence of coronary disease without symptoms CHF Hospitalization for CHF within past 6 mo Recent CABG, PTCA, or coronary stent ( 6 mo ago) Hospitalization for CHF 6 moago CHF with dyspnea that limits activities Angina not requiring hospitalization CABG, PTCA, or coronary stent ( 6 mo ago) CHF with dyspnea that has responded to treatment Exertional dyspnea or PND Sick sinus syndrome Arrhythmias Ventricular arrhythmia within the past 6 mo Ventricular arrhythmia 6 mo ago Chronic atrial fibrillation or flutter Pacemaker Hypertension DBP 130 mm Hg DBP mm Hg DBP mm Hg Severe malignant Secondary cardiovascular symptoms: papilledema or other eye vertigo, epistaxis, headaches changes DBP 90 mm Hg while taking antihypertensive medications Encephalopathy Venous disease PE 6 mo ago DVT controlled with coumarin or heparin Old DVT no longer treated with coumarin or Use of venous filters for PEs PE 6 moago heparin Peripheral arterial disease Bypass or amputation for gangrene or arterial insufficiency 6 moago Bypass or amputation for gangrene or arterial insufficiency 6 moago Chronic arterial insufficiency Intermittent claudication Untreated thoracic or abdominal aneurysm 6 cm Untreated thoracic or abdominal aneurysm 6 cm Status postabdominal or thoracic aortic aneurysm repair Abbreviations: CABG, coronary artery bypass graft; CHF, congestive heart failure; DBP, diastolic blood pressure; DVT, deep venous thrombosis; ECG, electrocardiogram; MI, myocardial infarct; PE, pulmonary embolism; PND, paroxysmal nocturnal dyspnea; PTCA, percutaneous transluminal coronary angioplasty. *Adapted from Piccirillo 2 with permission. For a complete form, see Table 2. American Society of Anesthesiologists Risk Classification System Risk Class Physical Status 1 Normal healthy patient 2 Patient with mild systemic disease 3 Patient with severe systemic disease 4 Patient with severe systemic disease that is a constant threat to life 5 Moribund patient who is not expected to survive without the operation diagnosed at the department of Otolaryngology Head and Neck Surgery of the Leiden University Medical Center between January 1999 and January All tumors were histologically confirmed SCC. Patients receiving palliative procedures (n=3) were excluded. The tumor sites were coded according the International Classification of Diseases for Oncology and included sinus, lip, oral cavity, nasopharynx, oropharynx, hypopharynx, and larynx. Age at diagnosis, sex, TNM stage, tumor site, prior malignancies, treatment, and follow-up data were obtained from the oncology database of our hospital, ONCDOC, which is maintained by specialized oncologic data managers. 4 Information concerning the surgical procedure, type of neck dissection, type of reconstruction, duration of anesthesia, and length of hospitalization were obtained from our hospital databases. Information about weight loss in the 6-month period prior to diagnosis and preoperative hemoglobin levels were determined from the patients files. A low hemoglobin level was defined as less than 13.0 g/l for women and less than 14.5 g/dl for men. 5 STUDY PROCEDURES Comorbidity and occurrence of major complications were determined from the patient s file. The ASA class, assigned by the attending anesthesiologist, was obtained from the original anesthesia form or from the hospital database. Each index was divided into 3 categories: ACE-27 grades were categorized as 0, 1, or 2 or higher; ASA classes were categorized as 1, 2, or 3 or higher. Major complications were defined according to the system outlined by Farwell et al 6 (Table 3) and recorded from the start of anesthesia until hospital discharge. Since all complications are recorded prospectively in the patients files and correspondence, the patients files could be used for this purpose. Medical and surgical complications were joined into a single variable, major complications, which was used in all analyses. One- and 6-month mortality rates were calculated. All statistical analyses were performed with SPSS 10.0 for Windows (SPSS Inc, Chicago, Ill). The impact of potential predictors on complications, mortality, and the length of hospitalization was analyzed univariately and multivariately. Univariate analyses were performed using the Pearson 2 test. A threshold of P.05 was set for entering multivariate analysis. Multivariate analyses were performed using binary logistic backward regression in all analysis. In cases where both the ACE-27 grade and ASA class entered a multivariate analysis, they were analyzed together as well as sepa- 28

3 rately to illustrate their individual strength. Only P values of.05 or lower were considered significant. RESULTS DEMOGRAPHICS The study population consisted of 120 surgically treated patients with HNSCC. Three patients were excluded because they received palliative procedures only; these 3 patients died within 1 month. The demographic characteristics of the remaining 117 patients are listed in Table 3. Most patients were men (63.2%) and older than 60 years (60.7%). The predominant tumor site was the oral cavity (50.4%). About 40% of all patients were staged T4; most patients were N0 (65.0%) and M0 (98.3%). A low hemoglobin level was found in 54 patients (49.5%). Eleven patients (9.6%) had lost more than 10% of their body weight in the 6 months prior to diagnosis. Fifty-five patients (47.0%) received local surgery only (including laser surgery). In 31 patients (26.5%), the surgical procedure consisted of a combined mandibular approach or a hemimaxillectomy. Twenty-four patients (20.5%) were treated with a laryngectomy. In 7 patients (6.0%), the surgical procedure was a neck dissection only. In a total of 61.5% of all cases, a neck dissection was performed. Anesthesia lasted over 6 hours in 52 patients (44.4%). Hospitalization lasted less than 2 weeks in most patients (60.7%). Most patients (61.6%) received postoperative radiotherapy (67 patients) or chemotherapy (3 patients) or both (2 patients). COMORBIDITY In all 117 patients, an ACE-27 grade could be assigned. Most patients (70.9%) received an ACE-27 grade of 0 or 1. Most patients (53.0%) were ASA class 2 (Table 3); in 3 patients (2.5%), local anesthesia was used; consequently, no ASA class was designated. Eleven patients (9.4%) had been diagnosed as having a prior malignancy. MAJOR COMPLICATIONS Twenty-two patients (18.8%) developed a major medical complication (Table 4); major surgical complications occurred in 10 patients (8.5%). Seven patients had both medical and surgical complications. Consequently, major complications occurred in 25 patients (21.4%). The mean age at diagnosis of patients with and without complications was 67.3 and 62.4 years, respectively. PREDICTION OF MAJOR COMPLICATIONS The univariate impact of all variables on major complications is detailed in Table 5. Because both ACE-27 grade and ASA class were significant predictors in univariate analysis, 3 multivariate analyses were performed. The first 2 multivariate analyses contained all significant variables but only 1 of the 2 comorbidity scales (ACE-27 grade or ASA class). The third multivariate analysis contained Table 3. Demographic Characteristics of Patients Undergoing Head and Neck Surgery Characteristic Category No. (%) of Patients (n = 117) Age at diagnosis, y 0 to (16.2) 50 to (23.1) (32.5) (28.2) Sex Female 43 (36.8) Male 74 (63.2) T stage 1 45 (38.5) 2 16 (13.7) 3 9 (7.7) 4 45 (38.5) X 2 (1.7) N stage 0 76 (65.0) 1 9 (7.7) 2 26 (22.2) 3 5 (4.3) X 1 (0.9) M stage (98.3) X 2 (1.7) Cancer site Lip 3 (2.6) Oral cavity 59 (50.4) Oropharynx 12 (10.3) Nasopharynx 1 (0.9) Hypopharynx 9 (7.7) Glottic larynx 15 (12.8) Supraglottic larynx 16 (13.7) Sinus 2 (1.7) Previous malignancies Yes 11 (9.4) No 106 (90.6) ASA class 1 28 (23.9) 2 62 (53.0) 3 24 (20.5) No ASA class assigned 3 (2.6) ACE-27 grade 0 48 (41.0) 1 35 (29.9) 2 34 (29.1) Hemoglobin level, g/l Female (50.5) Male 14.5 Female (49.5) Male 14.5 Missing 10 Weight loss, % (90.6) (9.6) Surgical procedure Local excision/laser therapy 55 (47.0) Laryngectomy 24 (20.5) Combined mandibular 31 (26.5) approach/ hemimaxillectomy Only neck dissection 7 (6.0) Neck dissection type No dissection 45 (38.5) Unilateral dissection 53 (45.3) Bilateral dissection 19 (16.2) Reconstruction type No reconstruction 79 (67.5) Free flap 17 (14.5) Myocutaneous flap 16 (13.7) Digestive tract 5 (4.3) Duration of anesthesia 6 h 65 (55.6) 6 h 52 (44.4) Length of 14 d 71 (60.7) hospitalization 14 d 46 (39.3) Adjunctive treatment Radiotherapy 67 (57.3) Chemotherapy 3 (2.6) Radiotherapy and 2 (1.7) chemotherapy None 45 (38.5) Death within 6 mo Yes 4 (3.4) No 113 (96.6) Abbreviations: ACE-27, Adult Comorbidity Evaluation 27 index; ASA, American Society of Anesthesiologists risk classification system. 29

4 Table 4. Major Medical and Surgical Perioperative Complications* No. (%) of Patients With Major Complications Perioperative Complications (n = 117) Total serious medical complications 22 (18.8) Cardiovascular (total) 7 (6.0) Arrhythmia 3 (2.6) Myocardial ischemia Myocardial infarction 1 (0.9) Congestive failure 2 (1.7) Code blue Pulmonary (total) 17 (14.5) Hypoxia 2 (1.7) Ventilator support 24 h 1 (0.9) Pneumonia 9 (7.7) Adult respiratory distress syndrome Bronchospasm Pulmonary embolism 1 (0.9) Other pulmonary 1 (0.9) Neurologic (total) 7 (6.0) Delirium 4 (3.4) Other neurologic 3 (2.6) Infectious (total serious) 12 (10.3) Surgical site infection deep 2 (1.7) Bacteremia Abscess 6 (5.1) Sepsis 2 (1.7) Other infectious 2 (1.7) Miscellaneous (total) Deep venous thrombosis Renal insufficiency Alcohol withdrawal Fall Other miscellaneous Unexpected transfer Death Total serious surgical complications 10 (8.5) Wound breakdown 1 (0.9) Fistula formation 5 (4.3) Flap donor and recipient site 4 (3.4) complications and failure Wound hematomas Need for additional unexpected procedure 5 (4.3) Total major complications 25 (21.4) Abbreviation:, did not occur in any patient. *Complications as defined by Farwell et al. 6 all significant variables, including both ACE-27 grade and ASA class. In the first multivariate analysis (Table 6), ACE-27 grade was the only remaining significant predictor of major complications. The odds ratios (ORs) and 95% confidence intervals (CIs) associated with ACE-27 grades were as follows: grade 1, 1.93 ( ); grade 2 or higher, 4.61 ( ). In the second multivariate analysis (Table 7), the duration of anesthesia and ASA class were significant predictors of major complications. The ORs (95% CIs) were as follows: duration of anesthesia greater than 6 hours, 7.75 ( ); ASA class 2, 2.04 ( ); and ASA class 3 or higher, 9.99 ( ). In the third multivariate analysis (Table 8), which included both ACE-27 grade and ASA class, ASA class had more impact on complications than did ACE-27 grade. The ORs (95% CIs) were as follows: duration of anesthesia greater than 6 hours, 7.75 ( ); ASA Table 5. The Impact on Major Complications of All Significant Variables in Univariate Analysis Variable Pearson 2 Value P Value ACE-27 grade ASA class T stage Surgical procedure Type of neck dissection Duration of anesthesia Abbreviations: See Table 3. Table 6. The Impact on Major Complications of All Significant Variables Excluding ASA Class in Multivariate (Backward Selection) Analysis ACE NA 0 NA NA 1.93 ( ) 2 NA 4.61 ( ) Abbreviations: ACE-27, Adult Comorbidity Evaluation 27 index; ASA, American Society of Anesthesiologists risk classification system; CI, confidence interval; NA, not applicable; OR, odds radio. Table 7. The Impact on Major Complications of All Significant Variables Excluding ACE-27 Grade in Multivariate (Backward Selection) Analysis ASA Class.003 NA 1 NA NA 2.04 ( ) 3 NA 9.99 ( ) Anesthesia duration, h.002 NA 6 NA NA 7.75 ( ) Abbreviations: See Table 6. class 2, 2.04 ( ); and ASA class 3 or higher, 9.99 ( ). PREDICTION OF MORTALITY None of the 117 patients died within 1 month after surgery, while 4 patients (3.4%) died within 6 months after diagnosis. In the univariate analysis, cancer site ( 2 =19.1; P=.01), hemoglobin level ( 2 = 4.2; P=.04), and postoperative adjunctive treatment ( 2 = 9.1; P=.03) were related with 6-month mortality. PREDICTION OF PROLONGED HOSPITALIZATION Ten variables were significant predictors of prolonged hospitalization in univariate analyses (Table 9). In the mul tivariate analysis, surgical procedure, type of neck dissection, and major complications remained significant (Table 10). 30

5 COMMENT Prevention of complications and mortality is an important issue in surgery. Remarkably, little is known about which factors contribute to the risk of complications and mortality. In this study, we chose to include a variety of potential risk factors: age at diagnosis, sex, site of the primary tumor, TNM stage, prior malignancies, weight loss, hemoglobin level, type of surgery, type of neck dissection, type of reconstruction, duration of anesthesia, postoperative adjunctive treatment, and comorbidity. Comorbidity has been established as an important factor in patients with head and neck cancer, and there are many indexes that reflect a patient s comorbidity. We chose the ACE-27 index because of its comprehensiveness. It is a complete review of all systems and easily applied, although it can be time-consuming. In a study by Piccirillo et al, 7 5 certified tumor registrars responded that coding comorbidity with a precursor of the ACE-27 index was not at all difficult or slightly difficult. These registrars needed from 1 to 15 minutes to abstract comorbidity information from a single medical chart. In a study by Paleri and Wight 8 in the United Kingdom, the ACE-27 grade could be successfully obtained by retrospective notes review. These researchers also made some suggestions to speed up the scoring process for the ACE-27 index. The ASA system was chosen because of its ready availability. It is a standard anesthesiologic procedure to designate an ASA class prior to every surgical procedure. Although the ASA class represents a patient s physical status, it was not designed for comorbidity measurements but to assess risks of anesthesia. In our study population, the incidence of major complications was 21.4%. This is in line with other studies. 6,9 In a study by Farwell et al, 6 34% of all patients had postoperative major complications in a study population of 93 patients with significant medical problems. However, both cancer and noncancer patients were included. In multivariate analysis, an anesthesia time of 8 hours or more was the only independent predictor of major complications. In a study by Pelczar et al 9 of 119 patients undergoing head and neck cancer surgery, 24% had at least 1 postoperative medical complication. The ASA class was a strong predictor of the development of postoperative medical complications in a univariate analysis (P.001). Both studies are in line with our results, which showed that both comorbidity scales and duration of anesthesia proved to be independent predictors of perioperative complications in multivariate analysis. In the present work and in the 2 above-mentioned studies, the number of patients was too low to allow for identification of single diseases contributing to the risk of complications. Therefore, correction of predisposing factors is not yet possible. However, it appears that it is worthwhile to investigate this further. In the meantime, a high comorbidity score should lead to a higher degree of caution by the treating physicians. In the present study, we found a 1-month mortality rate of 0%. This is in line with the results of Farwell et al, 6 who reported no deaths perioperatively in a population with Table 8. The Impact on Major Complications of All Significant Variables With ACE-27 Grade and ASA Class in Multivariate (Backward Selection) Analysis ASA Class.003 NA 1 NA NA 2.04 ( ) 3 NA 9.99 ( ) Anesthesia duration, h.002 NA 6 NA NA 7.75 ( ) Abbreviations: See Table 6. Table 9. The Impact on Length of Hospitalization of Variables With a Significant Impact in Univariate Analysis Variable Pearson 2 Value P Value ASA class T stage N stage Cancer site Weight loss Surgical procedure Type of neck dissection Reconstruction type Duration of anesthesia Major complications Abbreviation: ASA, American Society of Anesthesiologists risk classification system. Table 10. The Impact on Hospitalization of All Significant Variables in Multivariate (Backward Selection) Analysis* Surgical procedure.01 NA Local excision/laser NA 1.00 Laryngectomy NA ( ) Combined mandibular NA ( ) approach/hemimaxillectomy Neck dissection.04 NA No dissection NA 1.00 Unilateral dissection NA ( ) Bilateral dissection NA ( ) Major complications.02 NA No NA 1.00 Yes NA ( ) Abbreviations: CI, confidence interval; NA, not applicable; OR, odds ratio. *Patients with neck dissection only were not included in this analysis because all 7 of them left the hospital within 14 days. significant comorbidity. They propose that their improved outcome was owing to a team of attending internists who assisted them in the preoperative workup and through the perioperative period. Our patients were sent to an attending internist on indication only. Considering that all of our patients had cancer and a mortality rate of 0%, it is questionable whether the standard assistance of an internist was responsible for the decrease in mortality. 31

6 In a study by Bhattacharyya and Fried, 10 the mortality rate during hospitalization was 2.11%. This was determined in 3039 patients who underwent head and neck surgery, with a mean length of hospitalization of 6.2 days. We were also interested in the 6-month mortality rate because patients dying so soon after surgery would appear to have benefited very little from their treatment. Our 6-month mortality rate was 3.5%. In univariate analysis, 3 significant risk factors emerged: site of the primary tumor, adjunctive treatment, and a low hemoglobin level. All 4 patients who died in the 6-month period after diagnosis had a low hemoglobin level. The cause of death was residual or recurrent disease in 2 patients and cardiac disease in the other 2. In the first 2 patients, their anemia may have limited the effect of postoperative irradiation. 11 In the other 2 deaths, anemia may have played a role in the cause and severity of their cardiac disease. It seems worthwhile to study the effect of anemia correction in patients undergoing head and neck surgery, analogous to what has been done in studies of patients treated with irradiation. 12 It is striking that a low hemoglobin level (anemia) prior to surgery is not included in the evaluative criteria of the ACE-27 index. Anemia is an important medical condition, and it seems justified to include this important clinical finding in a comorbidity index. It is also striking that we did not find a relation between age and the development of major complications, nor did we find a relation between age and 6-month mortality. This implies that an equal treatment for all ages is justifiable. Another factor we studied was the length of hospitalization. This is an important factor with respect to patients well-being and the increasing costs of health care hospitalization, which must be reduced to the minimum. In the present study, hospitalization longer than 14 days was predicted by surgical procedure, type of neck dissection, and major complications. This stresses once more the need for further research to identify single diseases responsible for the development of major complications. In conclusion, duration of anesthesia and comorbidity as reflected by ACE-27 grade and ASA class are reliable predictors of major complications. Further research is needed to identify single diseases responsible for major complications to improve the physical condition and to reduce major complications, mortality rate, and length of hospitalization. Anemia should be incorporated in the ACE-27 system criteria, and a correction of low hemoglobin levels prior to surgery should be studied as a possible factor in reduction of the 6-month mortality rate. Submitted for Publication: April 17, 2003; final revision received August 31, 2004; accepted September 16, Correspondence: Marciano B. Ferrier, MD, Leiden University Medical Center, Department of Otolaryngology Head and Neck Surgery, PO Box 9600, 2300 RC Leiden, the Netherlands (M.B.Ferrier@LUMC.nl). Previous Presentation: This study was presented at the American Head and Neck Society Annual Meeting; May 4-6, 2003; Nashville, Tenn. REFERENCES 1. Piccirillo JF, Feinstein AR. Clinical symptoms and comorbidity: significance for the prognostic classification of cancer. Cancer. 1996;77: Piccirillo JF. Importance of comorbidity in head and neck cancer. Laryngoscope. 2000;110: Kaplan MH, Feinstein AR. The importance of classifying initial comorbidity in evaluating the outcome of diabetes mellitus. J Chronic Dis. 1974;27: Baatenburg de Jong RJ, Hermans J, Molenaar J, Briaire JJ, Le Cessie S. Prediction of survival in patients with head and neck cancer. Head Neck. 2001;23: Overgaard J, Hansen HS, Overgaard M, et al. A randomized double blind phase III study of nimorazole as a hypoxic radiosensitizer of primary radiotherapy in supraglottic larynx and pharynx carcinoma: results of the Danish Head and Neck Cancer study (DAHANCA) protocol Radiother Oncol. 1998;46: Farwell DG, Reilly DF, Weymuller EA, et al. Predictors of perioperative complications in head and neck patients. Arch Otolaryngol Head Neck Surg. 2002; 128: Piccirillo JF, Creech C, Zequeira R, et al. Inclusion of comorbidity into oncology data registries. J Registry Manage. 1999;26: Paleri V, Wight RG. Applicability of the Adult Comorbidity Evaluation-27 and the Charlson indexes to assess comorbidity by notes extraction in a cohort of United Kingdom patients with head and neck cancer: a retrospective study. J Laryngol Otol. 2002;116: Pelczar BT, Weed HG, Schuller DE, et al. Identifying high-risk patients before head and neck oncologic surgery. Arch Otolaryngol Head Neck Surg. 1993;119: Bhattacharyya N, Fried MP. Benchmarks for mortality, morbidity and length of stay for head and neck surgical procedures. Arch Otolaryngol Head Neck Surg. 2001;127: Bryne M, Eide GE, Lilleng R, et al. A multivariate study of the prognosis of oral squamous cell carcinomas. Cancer. 1991;68: Glaser CM, Millesi W, Kornek GV, et al. Impact of hemoglobin level and use of recombinant erythropoietin on efficacy of preoperative chemoradiation therapy for squamous cell carcinoma of the oral cavity and oropharynx. Int J Radiat Oncol Biol Phys. 2001;50: