Ann Otol Rhinal Laryngol101:1992

COMPLICATIONS OF PERCUTANEOUS ENDOSCOPIC GASTROSTOMY IN HEAD AND NECK CANCER PATIENTS SHARON BARRY

L.

WENIG,

E.

GIBSON,

MD

MD JAMES

L. WATKINS, MD

CHICAGO, ILLINOIS

Percutaneous endoscopic gastrostomy (PEG) has been shown to benefit patients with resectable carcinoma of the head and neck. In order to determine whether patients with existing tumor or postresection anatomic changes of the upper respiratory tract can undergo this procedure with an acceptably low complication rate, 349 patients with attempted PEG were studied. The PEG procedure was successful in 114 of 122 carcinoma patients, as compared to 220 of 227 patients in a control group (patients with neurologic disease). Intraoperative complications preventing PEG placement included pharyngeal or esophageal obstruction, inadequate transillumination of the abdominal wall, and respiratory distress and occurred in 7% of carcinoma patients and 3% of controls. The incidence of airway obstruction during endoscopy was equal between groups (1%). Postoperative complications related to the gastrostomy tube were more frequent in the nonhead and neck cancer group (14 % versus 5 %). Younger age, fewer concomitant medical problems, and better nutritional status may account for this difference. These findings suggest that preoperative, postoperative, and unresectable head and neck cancer patients are appropriate candidates for PEG, and postgastrostomy performance appears superior to that in other patient populations. KEY WORDS -

complications, head and neck cancer, percutaneous endoscopic gastrostomy.

INTRODUCTION

population that may account for this disparity.

Percutaneous endoscopic gastrostomy (PEG) has gained wide acceptance as a relatively safe and efficient means of enteral access. 1-4 Its use in the management of head and neck carcinoma has become a recent focus of attention in the otolaryngology literature.":" While most agree that PEG is beneficial for selected cancer patients, opponents to its routine placement fear an unacceptably high complication rate, resulting particularly from impedance by a tumor mass and respiratory distress during endoscopic insertion. 6 Postgastrostomy complications, especially in patients receiving concurrent chemotherapy, have been reported as high as 16% .2.4.7

MATERIALS AND METHODS

A retrospective analysis was performed of all patients undergoing PEG at the University of Illinois College of Medicine and affiliated hospitals between July 1988 and March 1991. Inpatient charts, tumor registry, and endoscopic reports were searched to document the primary diagnosis (reason for PEG), demographic characteristics, and outcome (success or failure to accomplish PEG). Concomitant medical disease was assessed by the number of failed or significantly impaired organ systems (cardiovascular, pulmonary, renal, hepatic, hematopoietic, or endocrinologic). Laboratory estimation of nutritional status by serum albumin and hematocrit at the time of PEG was recorded. Patients with a diagnosis of squamous cell carcinoma of the head and neck were categorized by primary site, tumor stage, and disease status. Patients' disease status was designated as preoperative resectable, postoperative with no evidence of active disease, or unresectable. Previous or concomitant irradiation or chemotherapy, and its timing relative to PEG, was noted.

Our recent study of preoperative PEG in resectable carcinoma patients did not substantiate these concerns." We observed no instances of airway compromise during 43 procedures and had a tube-related complication rate of less than 5 %. These patients, however, comprised a select group of patients with primary, resectable malignancies of the head and neck. This study was undertaken to determine whether candidacy for PEG applies to all patients with carcinoma of the head and neck, including those awaiting resection and those who have had resection as well as those with unresectable disease. Candidacy is based on an acceptable complication rate as compared with a control group with normal head and neck anatomy. As the literature to date demonstrates a surprisingly low postgastrostomy complication rate for head and neck cancer patients, we will also explore patient characteristics within each

Patients undergoing PEG for non-head and neck carcinoma diagnoses comprised the control group. These included patients with inadequate oral alimentation due to altered mental status, chronic aspiration of central neurologic origin, neuromuscular dysphagia, and cachexia associated with disseminated nonsquamous malignancies. The PEG was performed in all patients under 10-

From the Department of Otolaryngology-Head and Neck Surgery (Gibson, Wenig) and the Division of Gastroenterology, Department of Internal Medicine (Watkins), The University of Illinois College of Medicine at Chicago, Chicago, Illinois. Presented at the meeting of the American Broncho-Esophagological Association, Waikoloa, Hawaii, May 5-6, 1991. REPRINTS - Barry L. Wenig, MD, 1855 W Taylor St, Suite 2.42, Chicago, IL 60612.

46

Downloaded from aor.sagepub.com at The University of Auckland Library on June 10, 2015

47

Gibson et al, Percutaneous Endoscopic Gastrostomy TABLE 1. DISTRIBUTION OF HEAD AND NECK CARCINOMA PATIENTS BY SITE AND STAGE

Stage II Stage III Stage IV Total

Larynx

Tongue

Tonsil

Oral CaVity

Total

12 10 17 39

2 7 14 23

4 7 19 30

0 11 11 22

18 35 61 114

cal anesthesia with intravenous sedation (0.5 mg of Versed and 30 to 50 mg of Demerol). The operative procedure is performed as follows. With the patient in the supine position, a flexible fiberoptic gastroscope is introduced perorally and advanced to the region of the gastric fundus. Transillumination of the anterior abdominal wall by the gastroscope directs the point of sterile puncture. A silk suture is passed through the puncture site and grasped by the endoscope, which is then brought out through the mouth. A 16F to 22F rubber gastrostomy catheter is secured to the orally extruded suture and pulled into the stomach by withdrawal of the suture through the abdominal wall. The tube is stabilized by internal and external bolsters. Feedings are begun following the return of bowel function, generally within 24 hours of placement. Failure to accomplish PEG within each group was documented. Reason for failure was attributed to pharyngeal or esophageal luminal obstruction preventing insertion of the gastroscope, inadequate transillumination of the abdominal wall to allow safe puncture, or intraoperative respiratory distress. Gastrostomy-related complications, ie, those related to the indwelling catheter, were sought and documented. Possible unfavorable events included major (sepsis, pulmonary aspiration, abdominal wall breakdown) as well as minor complications (gastroparesis or ileus, cellulitis, chronic gastric leakage). RESULTS

During the 32-month study period, PEG was attempted with 349 patients. Successful placement was achieved in 334, of whom 114 had existing or previous squamous cell carcinoma of the head and neck. The remaining 220 required PEG for primarily neurologic deficits. The distribution by primary site and stage" of TABLE 2. REASONS FOR FAILURE TO ACCOMPLISH PERCUTANEOUS ENDOSCOPIC GASTROSTOMY

Pharyngeal or esophageal obstruction Inade:huate transillumination of a dominal wall Intraoperative respiratory distress Total

Carcinoma Group (n = 114)

Neurologic Group (n = 220)

3 (2.6%)

1 (0.5%)

4 (3.5%)

4(1.8%)

1 (0.9%) 8 (7.0%)

2 (0.9%) 7 (3.2%)

TABLE 3. POSTOPERATIVE GASTROSTOMY COMPLICATIONS IN HEAD AND NECK CARCINOMA GROUP Sepsis Breakdown of abdominal wound Ileus Cellulitis Total (n = 114)

No. of Pts

%

1

0.9

1 1 3 6

0.9 0.9 2.6 5.2

carcinoma patients with successful PEG is shown in Table 1. Of this group, 78 underwent PEG prior to primary resection, 13 postresection with no evidence of tumor, and 23 with unresectable disease. Twenty-six patients had received full-course primary or postoperative radiotherapy prior to PEG. Chemotherapy consisting of 5-fluorouracil and cisplatin for two or three cycles was administered within 3 months of PEG in 15 patients. Three received chemotherapy prior to PEG; 12 began chemotherapy immediately afterward. Failure to accomplish PEG occurred in eight patients in the head and neck carcinoma group (7 %) and seven in the neurologic group (3 %). Reasons for failure in each group are listed in Table 2. The incidence of intraoperative respiratory distress was equal between groups at 1 % . Following successful PEG, six head and neck cancer patients (5 %) experienced gastrostomy-related complications (Table 3). No cases of aspiration or hemorrhage were observed. The patient in which abdominal wound breakdown occurred had completed three cycles of chemotherapy 30 days before PEG insertion. Postoperative complications related to gastrostomy occurred in 31 patients (14 %) of the neurologic group, as listed in Table 4. Massive pulmonary aspiration was seen in 3 patients, resulting in respiratory failure in all 3 and death in 1. Leakage of gastric contents, which was not seen in the carcinoma group, was chronic in 15 neurologic patients. Patient characteristics with possible bearing on gastrostomy-related complications are summarized in Table 5. Demographics include an age range for the carcinoma patients of 33 to 87 and a range of 41 to 90 for the neurologic group. The male-to-female ratio was approximately 5: 1 for head and neck canTABLE 4. POSTOPERATIVE GASTROSTOMY COMPLICATIONS IN NEUROLOGIC GROUP No. of Pts

Sepsis Massive aspiration Ileus Breakdown of abdominal wound Cellulitis Gastric leakage Total (n = 220)

Downloaded from aor.sagepub.com at The University of Auckland Library on June 10, 2015

%

1 3 2

0.5 1.4 0.9

4 6 15 31

1.8 2.7 6.8 14

Gibson et al, Percutaneous Endoscopic Gastrostomy

48

TABLE 5. PATIENT CHARACTERISTICS

Carcinoma Group Average age (y) No. of organ systems compromised Hematocrit (%) Serum albumin (g/dL)

Neurologic Group

55

68

1 37.8 3.5

3 28.2 3.0

cer patients and 7:3 for the control group. Serum albumin (normal values, 4.0 to 5.5) and hematocrit (normal, 40.0 to 50.0) were decreased in both groups, but to a greater degree in the neurologic group. Statistical significance between the two groups, however, was not reached for either variable. DISCUSSION

Gastrostomy has become the preferred means of alimentation for patients with chronic dysphagia and aspiration. The Stamm gastrostomy, in which concentric pursestring sutures are placed through the adjacent gastric musculature and mucosa, was introduced in 1894 10 and remains in use today. Associated problems including gastric leakage and tube dislodgment led to the development of permanent mucosal-lined gastrostomies, of which many modifications have been described. II Reported complication rates with these procedures range from 15 % 12 to 24 % .13 Major morbidity includes wound dehiscence, peritonitis, and hemorrhage and is largely related to the laparotomy required for these gastrostomies. The need for an open abdominal procedure has been virtually eliminated by the development of percutaneous gastrostomy placed via endoscopic visualization. I The PEG is performed routinely under local anesthesia. It is easily tolerated by patients with normal aerodigestive tract anatomy and carries an acceptably low risk of intraoperative complications in this population. 14 For patients with head and neck carcinoma, the ability to undergo PEG and its associated risks is of special interest. The distal gastrointestinal tract is usually intact and therefore amenable to enteral feedings. Dysphagia and aspiration from an existing tumor mass, postresection anatomic alteration, or mucositis from adjuvant irradiation and chemotherapeutic regimens frequently prevent oral alimentation. The relative comfort, ease of management, and availability for long-term or intermittent use make gastrostomy a desirable means of enteral access for these patients. In addition, rehabilitation of speech and deglutition can proceed without the hindrance of a nasogastric tube and with nutritional supplementation as needed. It is therefore important to determine whether patients with head and neck cancer are good candidates for PEG. Postgastrostomy problems with gastroparesis,

chronic leakage, abdominal wall cellulitis, or breakdown and sepsis have occurred in up to 16 % of patients following PEG performed for neurologic debilitatlon.v":" Pulmonary aspiration after PEG has been reported in 36% of neurologically compromised patients," carrying a significant risk of mortality. In contrast, the observed gastrostomy-related complication rate in our initial study of patients undergoing primary resection of head and neck carcinoma was less than 5 % .8 Shike et al" found a similarly low incidence in their cancer population in which the mean length of gastrostomy-dependent alimentation was 135 days. One case of aspiration pneumonitis was noted in their series. This disparity of complication rates between the head and neck cancer patients and those with neurologic indications for PEG is reaffirmed in the current study. Five percent of the carcinoma group experienced post gastrostomy problems. Half of these had local cellulitis at the puncture site that resolved with enteral antibiotic therapy. Only 1 patient developed a major complication, sepsis, requiring parenteral antibiotics and feedings. Similar complications were seen in the neurologic group, but at a significantly higher rate (14 %). In addition, massive aspiration with subsequent pulmonary failure occurred in 3 patients. Chronic gastric leakage, a complication not observed in the head and neck cancer patients, was recorded in 15 patients in the neurologic group. These differences may be based on the more debilitated state of patients with advanced neurologic disease. The number of accompanying failed organ systems parallels the increased age in the neurologic group. Patients with concomitant peripheral vascular disease and metabolic or endocrinologic aberrations (eg, diabetes mellitus) have impaired wound healing and may be predisposed to development of chronic leakage or cellulitis at the catheter site. Similarly, malnutrition contributes to poor tissue integrity, local wound breakdown, and infection. Although both patient groups demonstrated low hematocrit and serum albumin values, those in the neurologic group were more severely depressed. The incidence of gastrostomy-related complications in patients receiving chemotherapy is an important issue. Cytotoxic agents, often given in combination with radiotherapy, commonly cause mucositis of sufficient severity to prevent adequate oral nutrition, forcing patients to rely significantly on enteral supplements. Unfortunately, the fear of neutropenia and thrombocytopenia leading to bleeding, cellulitis, chronic drainage, or sepsis has prompted some authors to condemn PEG in the peri chemotherapeutic period. We found only 1 of 15 patients undergoing full-course chemotherapy within 3 months of PEG placement to develop wound breakdown, and it did not progress to sepsis and responded to local care. No episodes of temper-

Downloaded from aor.sagepub.com at The University of Auckland Library on June 10, 2015

Gibson et al, Percutaneous Endoscopic Gastrostomy

ature elevation, bleeding, or overt sepsis were noted in this group. On the basis of these observations, we cannot support the routine rejection of all patients on a chemotherapeutic protocol from candidacy for PEG. Further studies involving larger patient numbers and specific determination of platelet and white blood cell counts are needed to more accurately predict whether certain patients are at an unacceptably high risk for this procedure. In the analysis of unsuccessful attempts at PEG, the failure rate was higher for the head and neck cancer patients than those in the neurologic group (7 % versus 3 %). These rates are in agreement with those in the existing literature, in which PEG was unsuccessful in 4 % to 7 % of attempts for all patient populations.sv"" In both groups, inadequate transillumination of the abdominal wall was responsible for half of the placement failures. This was attributable to many factors, including anatomic alteration from previous gastrointestinal surgery, highriding stomach (beneath an elevated left hemidiaphragm), and morbid obesity. As other patients with similar features underwent PEG uneventfully, poor outcome cannot adequately be predicted preoperatively from these findings alone. Any contraindication to PEG in the head and neck cancer population therefore lies in the inability to successfully pass the gastroscope to the gastric fundus. This may arise in two situations: luminal narrowing with a diameter too small to accommodate the endoscope, and intraoperative respiratory distress. Obstruction of the upper aerodigestive tract may occur at the level of the oral cavity, oropharynx, hypopharynx, or esophagus and may be due to tumor bulk, irradiation fibrosis, or postoperative changes. Advanced laryngeal lesions may also obscure the cricopharyngeal inlet. In our study population, 61 of 114 patients had stage IV cancer, and only 1 with a tonsillar lesion involving the lateral and posterior walls failed PEG because of obstructing tumor. Another patient had a concomitant esophageal carcinoma that prevented passage of the gastroscope. Of 14 patients with surgical alteration of the head and neck, one had a prohibitive pharyngeal stricture. An additional 26 patients underwent PEG successfully following full-course radiotherapy; this finding suggests that irradiation-induced fibrosis does not generally hinder placement. Overall, 3 % of our carcinoma patients failed PEG because of oropharyngeal or esophageal luminal obstruction regardless of preoperative, postoperative, or unresectable disease status. Empirically, only patients with tumors involving the pterygoid muscula-

49

ture with severe trismus preventing peroral admittance of the endoscope should be routinely excluded as candidates for PEG. Intraoperative respiratory distress is a valid concern for head and neck cancer patients when instrumentation is performed in the presence of a marginal airway. The routine administration of narcotics and sedatives prior to endoscopy certainly predisposes to this complication. Hunter et al" advocate deferral of PEG until after tracheotomy or resection when managing posterior pharyngeal and pyriform sinus lesions. In this study, 9 carcinoma patients had a preexisting tracheotomy or laryngostoma at the time of PEG. In the remaining 105, PEG was attempted without a secured airway. This group included 19 patients with stage IV pharyngeal lesions and 13 with advanced (stage III or IV) pyriform tumors. One patient who had stage IV disease of the tonsil experienced respiratory distress intraoperatively. The risk of intraoperative complication with PEG for advanced lesions therefore does not appear prohibitive. In fact, airway compromise occurred with identical frequency in the neurologic group (1 %). Although no mass impingement on the respiratory tract occurs in this group, altered mental status with poor airway protection and impaired clearance of secretions lead to respiratory decompensation. Judicious administration of sedation is therefore crucial in all patients. CONCLUSIONS

Patients with preoperative or unresectable carcinoma of the head and neck, as well as those who have had resection, are candidates for PEG. The failure rate for endoscopic placement was 7 %, as compared to 3 % in controls undergoing PEG for neurologic disease. However, only half of the failures in the carcinoma group were due to pharyngeal or esophageal obstruction. The incidence of respiratory distress during PEG with advanced tumors did not exceed that of the control group. Head and neck cancer patients demonstrated a significantly lower rate of postgastrostomy complications (5 % versus 14 %), which may be attributed to younger age, better nutritional status, and fewer concomitant medical problems within this group. These findings indicate that the risk of complications with PEG is not prohibitive in head and neck cancer patients, and indeed they appear superior candidates for PEG as compared to our control group.

REFERENCES 1. Ponsky j L, Gauderer MWL. Percutaneous endoscopic gastrostomy: nonoperative technique for feeding gastrostomy. Gastrointest Endosc 1981;27:9-11. 2. Ponsky ]L, Gauderer MWL, Stellato T A, Aszodi A. Percu-

taneous approaches to enteral alimentation. Am Surg 1985;149: 102-5. 3. Miller RE, Kummer BA, Kotler DP, Tiszenkel HI. Percutaneous endoscopic gastrostomy: procedure of choice. Ann Surg

Downloaded from aor.sagepub.com at The University of Auckland Library on June 10, 2015

50

Gibson et al, Percutaneous Endoscopic Gastrostomy

1986;204:543-5. 4. Larsen DE, Burton DO, Schroeder KW, DiMagno EP. Percutaneous endoscopic gastrostomy: indications, success, complications and mortality in 314 consecutive patients. Gastroenterology 1987;93:48-52. 5. Shike M, Berner YN, Gerdes H, et al. Percutaneous endoscopic gastrostomy and jejunostomy for long-term feeding in patients with cancer of the head and neck. Otolaryngol Head Neck Surg 1989;101:549-54. 6. Hunter JG, Lauretano L, Shellito PC. Percutaneous endoscopic gastrostomy in head and neck cancer patients. Ann Surg 1989;210:42-6. 7. Stern JS. Comparison of percutaneous endoscopic gastrostomy with surgical gastrostomy at a community hospital. Am J Gastroenterol 1986;81: 1171-3. 8. Gibson SE, Wenig BL. Percutaneous endoscopic gastrostomy in the management of head and neck carcinoma. Laryngoscope (in press).

9. American Joint Committee on Cancer. Manual for staging of cancer. 3rd ed. Philadelphia, Pa: JB Lippincott, 1988. 10. Anderson P, Woodward ER. Gastrostomy. Am J Surg 1972;124:581-6. 11. Webster M, Carey L, Ravitch M. The permanent gastrostomy. Arch Surg 1975;110:658-60. 12. Wasiljew B, Ujiki G, Beal J. Feeding gastrostomy: complications and mortality. Am J Surg 1982;143:194-5. 13. Wilkinson W, Pickelman J. Feeding gastrostomy: a reappraisal. Am J Surg 1982;48:273-5. 14. Van der Werken C, van Vroonhoven TJMV, Juttman JR, Stuifbergen WNHM. Gastropexie in Kombination mit Percutaner Endoskopischer Gastrostomie. Chirurg 1987;58:118-9. 15. Burtch GO, Shatney CH. Feeding gastrostomy tant or assassin? Am Surg 1985;51 :204-7.

assis-

16. Kirby OF, Craig RM. Percutaneous endoscopic gastrostomies: a prospective evaluation and review of the literature. JPEN J Parenter Enteral Nutr 1986;10:155-9.

7TH WORLD CONGRESSES FOR BRONCHOLOGY & BRONCHOESOPHAGOLOGY The 7th World Congresses for Bronchology & Bronchoesophagology will be held Sept 28-0ct 2, 1992, at the Mayo Clinic and Mayo Medical Center in Rochester, Minnesota. For further information, contact Udaya B. S. Prakash, MD, Secretary-General & Director, 7th WCB & WCBE, East-18, Mayo Clinic, Rochester, MN 55905.

Downloaded from aor.sagepub.com at The University of Auckland Library on June 10, 2015

Complications of percutaneous endoscopic gastrostomy in head and neck cancer patients.

Percutaneous endoscopic gastrostomy (PEG) has been shown to benefit patients with resectable carcinoma of the head and neck. In order to determine whe...
475KB Sizes 0 Downloads 0 Views