Defining the Indications for Prophylactic Percutaneous Endoscopic Gastrostomy Tubes in Surgically Treated Head and Neck Cancer Patients

Malnutrition is a common problem in head and neck cancer with up to 50% of the patients developing some degree of nutritional deficiency.1 The etiologies of this problem can be divided into two categories: tumor related or treatment related. Patient with tumor related malnutrition typically present with obvious clinical signs and symptoms of undernourishment. Tumor cachexia can contribute but this is primarily caused by physical impediments to oral consumption such as pain, oropharygeal obstruction, or nerve compression , all resulting in discordant degluttination.2 Prior to definitive cancer therapy, this group of patients requires nutritional resuscitation. Additionally many patients who present nutritionally sound and who undergo surgical resection experience some degree of postoperative nutritional difficulties. In many cases it a short lived and inconsequential. However, a subset of patients will experience a more severe prolonged course requiring enteral supplementation. Prior studies have shown that if these patients are not supplemented, they will likely experience severe dehydration, treatment intolerance, or severe treatment related complications that require hospitalization.3,4 Inherently, the surgical treatment of head and neck malignancies can be quite debilitating and result in significant mastication and deglutination dysfunction.5-7 At times this involves radical resections that require complex reconstructions to maintain oropharyngeal continuity; and adjuvant radiation and/or chemotherapy may be required to maximize local control. Indiscriminately placing PEG tubes in all patients would put many patients at risk for PEG related complications.8-13 Therefore, the difficulty has been trying to preoperatively identify which patients likely to experience prolonged treatment induced malnutrition and benefit from early nutritional supplementation. Groups have identified some factors that are predictive of a need for enteral support.3,4,14-17 One factor that has been clearly established is radiation therapy, particular when given


Introduction
Malnutrition is a common problem in head and neck cancer with up to 50% of the patients developing some degree of nutritional deficiency. 1 The etiologies of this problem can be divided into two categories: tumor related or treatment related. Patient with tumor related malnutrition typically present with obvious clinical signs and symptoms of undernourishment. Tumor cachexia can contribute but this is primarily caused by physical impediments to oral consumption such as pain, oropharygeal obstruction, or nerve compression , all resulting in discordant degluttination. 2 Prior to definitive cancer therapy, this group of patients requires nutritional resuscitation. Additionally many patients who present nutritionally sound and who undergo surgical resection experience some degree of postoperative nutritional difficulties. In many cases it a short lived and inconsequential. However, a subset of patients will experience a more severe prolonged course requiring enteral supplementation. Prior studies have shown that if these patients are not supplemented, they will likely experience severe dehydration, treatment intolerance, or severe treatment related complications that require hospitalization. 3, 4 Inherently, the surgical treatment of head and neck malignancies can be quite debilitating and result in significant mastication and deglutination dysfunction. [5][6][7] At times this involves radical resections that require complex reconstructions to maintain oropharyngeal continuity; and adjuvant radiation and/or chemotherapy may be required to maximize local control. Indiscriminately placing PEG tubes in all patients would put many patients at risk for PEG related complications. [8][9][10][11][12][13] Therefore, the difficulty has been trying to preoperatively identify which patients likely to experience prolonged treatment induced malnutrition and benefit from early nutritional supplementation. Groups have identified some factors that are predictive of a need for enteral support. 3,4, [14][15][16][17] One factor that has been clearly established is radiation therapy, particular when given postoperatively. 3,4,17 Other factors such as Stage IV disease, base of tongue tumor location, and heavy alcohol ingestion are less clearly defined. [14][15][16] Some criticisms have been that the studies conducted to identify these variables were small, used a mixed population of surgical and non-surgical patients, and often used durations of enteral support that were short (< 4 weeks) or undefined. At Roswell Park Cancer Institute (RPCI) prophylactic PEG are routinely placed in surgically treated patients who require a composite resection, flap reconstruction, radiation therapy, chemotherapy, and at the discretion of the attending surgical staff. We found that many patients required their PEG tube for 4 weeks or less; while some patients required their PEG for a year or longer. In order to more accurately define which patients benefited from prophylactic PEG placement we reviewed our experience. We used a homogenous population of surgically treated head and neck cancer patients to identified patient, tumor, and treatment factors that were predictive of a short-term (≤ 3 months) and long-term (≥1 year) PEG tube dependency.

Methods
One hundred forty one cases of surgically treated head and neck cancers treated at RPCI from January 1, 1999 to December 31, 2003 who underwent pretreatment placement of PEG tubes were reviewed. Only patients with squamous cell carcinoma (SCC) of the oral cavity, oropharynx, larynx, and pharynx were included in this study eliminating 14 patients. Seven patients had PEG tube placed a second time for the treatment of a new primary or recurrent disease, and six patients did not have complete records, leaving 114 patients for evaluation. The variables analyzed were divided into patient factors (age and sex), tumor factors (primary site, T stage and nodal status), and treatment factors (flap reconstruction, radiation, and chemotherapy). A short-term dependency required that the PEG tube be in place for 3 months or less, while a long-term dependency require usage for a year or longer.

Statistical method
The duration of PEG tube dependency was calculated from the date of placement until the time of removal. Patients who had their PEG tube removed and not replaced were considered to be no longer dependent on it. Patients who died while still dependent on their PEG tube, or who were still dependent at last documented follow-up were considered to have censored durations. Because of this censoring, time to event analyses was used. The distribution of PEG duration was compared across age, sex, tumor sites, T stage, N stage, flap reconstruction, radiation, and chemotherapy. Kaplan-Meier estimates of the proportions (and 95% confidence intervals {CIs}) of patients with PEG tubes in place at 3 and 12 months after placement were determined for each variable and log-rank tests were used to compare durations. Proportional hazards regression models were used to compare durations while accounting for other factors. Variables were selected for inclusion in the model in a stepwise selection process. Variables were entered in the model if p<0.05 and were retained if p<0.05. Because a number of patients had unknown T stage (8 patients) and unknown N stage (20 patients), in the proportional hazards regression models, these factors included separate levels for www.intechopen.com 'unknown' T and 'unknown' N stage, respectively. In the log-rank analyses of T stages and N stages, patients with unknown stages were not included. These analyses are post hoc so no adjustments for multiplicity are made. All tests were done two-sided with a significance level of 0.05. All analyses were done using SAS version 8.2.

Results
The mean age of patients in this study was 65 and 65% of patients were male. Sixty-four percent of the patients had advanced T stage or recurrent disease and node positive disease was present in 49% of the patients. Flap reconstructions were performed in 39% of patients, while the rate of adjuvant therapy was 40% for radiation and 11% for chemotherapy. The percentage of patients in each of the four major tumor sites were 42% oral cavity, 23% oropharynx, 26% larynx, and 9% pharynx. In Table 1 the patient characteristics (age, sex), tumor characteristics (T stage, N stage), and treatment (flap reconstruction, radiation, chemotherapy) characteristic, along with PEG status at the end of the study is presented for each major tumor site. Overall 64% (73/114) of patients in the study had their PEG tubes removed. Sixty-nine percent of patients had short-term PEG usage. When the group receiving adjuvant radiation was compared to the group that did not receive radiation treatment a significant difference was observed 91% (83, 99) vs. 53% (41, 65). Eighty-nine percent of pharyngeal tumor site patient and 92% of chemotherapy patients had a short term dependency but this was not statistically different from the other tumors site or the no chemotherapy group respectively. The short-term dependency was not influenced by patient age, sex, T stage, N stage, or flap reconstruction ( Table 2). The long-term dependency for this group of patients was 36%. Table 2 presents Kaplan-Meier estimates of the proportions of patients with PEG tubes in place at 12 months for each variable. PEG tube duration was statistically significantly different across surgical sites: 78% of pharynx patients still had their tubes in place after 12 months, while only 45% of oral cavity patients, 34% of oropharynx patients, and 11% or larynx patients had tubes in place. Patients who underwent flap reconstruction also had statistically significantly (p=0.004) longer PEG tube durations than those who did not, 52% vs. 25%, respectively. When a multivariate analysis was performed, adjusting for other factors, site and flap reconstruction remained as statistically significantly long-term risk factors. Chemotherapy was also found to be significant factor in long-term PEG dependency.

Discussion
The prevention of malnutrition and early nutritional support in the management of cancer patients is well documented. Specifically in head and neck cancer, suboptimal nutrition during definitive treatment results in a significant increase in surgical complications, dehydration, therapy breaks and hospitalizations. (3). Because malnutrition can result in reduced immunosurvelliance, it may contribute to early local and distant cancer recurrence. (18,19) Avoidance of treatment induced malnutrition may prevent these complications, and prophylactic placement of PEG tube provides access for the delivery of nutrition. Identification of the risk factors which contribute to prolonged enteral support ensures PEG placement in the appropriate patients. Of equal importance is that accurate risk www.intechopen.com factors can avoids subjecting low risk patients to PEG related complications, and costs. In this study we found that radiation was the only predictor of a short-term dependency; while pharyngeal tumor site, flap reconstruction, and chemotherapy were predictive a long-term need. Other groups have documented that radiation treatment results in significant malnutrition. (3,4,20) Radiation treatment often results in xerostomia, loss of taste, mucositis, and tumor edema which all contribute to poor oral intake and nutritional deficits. When patients receive primary radiation without nutritional support 40% of patient loose more than 10% of their baseline weights, 40% of patients require hospitalization during the treatment with 20% of patients requiring therapy break, and 40% of these patients will require a gastrostomy placement to complete therapy. (3,4,20) As expected 91% of patients who received radiation in addition to surgical resection required short term enteral support. However, this dependency was self-limited, and did not impact long term oral nutrition, evident by no difference at 1 year between the radiated or unirradiated (Table 2). A pharyngeal site was significant on univariate analysis, while flap reconstruction was significant on both univariate and multivariate analysis. Given the significant amount of dysfunction associated with these surgical procedures, this data validates our clinical expectation that these subsets benefit from prophylactic PEGs. Importantly, given the propensity of oral pharyngeal bacterial overgrowth in this patient subset and significant intra-oral tumor burden, we believe that the T-fastener technique should be used to prevent/reduce PEG site abscess and local cancer recurrence. We previously published that the T-fastener technique has a low rate of local infection and cancer recurrence in head and neck carcinoma population. (21) Chemotherapy was the only other significant factor on multivariate analysis. Although our series was small, the increasing use of chemotherapy in the management of the head and neck SCC population will dramatically increase this patient fraction requiring nutritional support. In our study almost half of chemotherapy patients required long-term support. We believe that prophylactic PEG placement should be part of the management discussion in patients receiving adjuvant or neoadjuvant chemotherapy. Importantly when PEG tubes are placed in this patient population, the usage of the Tfastener technique is critical. Most of these patients are at high risk for PEG site infection and tumor implantation when the pull through technique is used in this patient population. We recently published that the rate of these complications can be significantly reduced by direct PEG placement with T-fastener strategy. This review confirmed the favorable current approach to prophylactic PEG tube in the head and neck cancer population. Aside from patients who present malnourished, prophylactic P E G t u b e s s h o u l d b e p l a c e d i n a l l S C C h e a d a n d n e c k c a n c e r p a t i e n t s w h o h a v e a pharyngeal primary tumor site, require flap reconstruction, undergo radiation therapy, and/or chemotherapy.