The role of radiotherapy and musculocutaneous flaps in oropharyngocutaneous fistulas

The Role of Radiotherapy and Musculocutaneous Flaps in Oropharyngocutaneous Fistulas Pierpont F. Brown III, MD, John J. Coleman III, MD, Atlanta,Georgia The development of oropharyngocutaneous fistulas after resection of the head and neck is a major cause of morbidity. In this retrospective review, we examined the role of radiotherapy and musculocutaneous flaps in the development and m a n a g e m e n t of fistulas and found several significant factors as-

sociated with these fistulas. In patients with this complication, there is a significant decrease in postoperative hemoglobin and albumin levels and absolute lymphocyte counts. In addition, there is a significant reduction in lymphocyte counts in the irradiated patient. Musculocutaneous flaps play an invaluable role in the m a n a g e m e n t of fistulas. In the irradiated patient, the healing of fistulas is a long and difficult process; the presence of a musculocutaneous flap is the only significant factor in the ability of the fistulas to heal spontaneously. If the fistulas are to close spontaneously, they will do so within 4 to 6 weeks. Fistulas that develop in the absence of a previously placed musculocutaneous flap are more severe than if a vascularized flap were present initially. Successful m a n a g e m e n t of such fistulas is enhanced by use of vascularized extracervical tissue.

From the Department of Surgery, Emory University School of Medicine, Atlanta, Georgia. Requests for reprints should be addressed to John J. Coleman III, MD, 25 Prescott Street, Northeast, Suite 3420, Atlanta, Georgia 30308. Presented at the 34th Annual Meeting of the Society of Head and Neck Surgeons, New Orleans, Louisiana, May 22-26, 1988.

256

THE AMERICAN JOURNAL OF SURGERY

he development of oropharyngocutaneous fistulas T after resection for head and neck carcinoma results in a prolonged hospital stay, delay in institution of adjuvant therapy, and a serious reduction in postoperative quality of life in patients whose late-stage disease demands an early return to function before recurrence proves fatal. Many factors have been implicated in the development of oropharyngocutaneous fistulas, resulting in ranges in incidence from 6 percent for oral resection to 38 percent for laryngectomy, and an overall incidence for major head and neck resections of about 13 percent [13]. In this report, we review the roles of radiotherapy and the musculocutaneous flap in the development and management of oropharyngocutaneous fistulas. MATERIAL AND METHODS A retrospective review identified 54 patients with 55 oropharyngocutaneous fistulas that developed after extensive resections of the head and neck at the Emory University affiliated hospitals from January 1982 to June 1987. The patient population was elderly (range 37 to 76 years, average age 60 years), predominately male (42 of 54 patients), and presented with late-stage disease. Of those patients who had a positive response to radiotherapy and could be staged, 25 of 30 presented with stage III or higher disease. All patients underwent resection for squamous cell carcinoma except for two who had radionecrosis after definitive radiotherapy, one with follicular carcinoma of the thyroid and one with pleomorphic adenoma of the salivary gland. Twenty-two of 54 patients underwent definitive radiotherapy an average of 28 months (range 6 to 168 months) prior to resection of the disease, receiving an average of 6,124 rads (16 of 22 patients). Two had previously been irradiated for breast cancer and one for thyroid cancer. Fifty-six percent of patients had oropharyngeal resections and 44 percent had laryngectomy or laryngopharyngectomy. Sixty-nine percent had concomitant radical neck dissections; of these, 32 percent had prior radiotherapy. Surgical defects were closed in a variety of fashions: by primary approximation (25 patients), pectoralis major musculocutaneous flap (20 patients), jejunal free flap (9 patients), or free radial forearm flap (1 patient). Although all patients were tobacco and alcohol users, liver disease, diabetes, peripheral vascular disease, and heart disease were not serious problems. All patients received perioperative antibiotics and enteral feeding. The patients were divided into four groups according to their radiation status and their index resection and reconstruction: preoperative radiotherapy with resection and primary closure (Group I), preoperative radiotherapy with resection and flap closure (Group II), primary closure with no preoperative radiotherapy or resection (Group III), or flap closure with preoperative radiothera-

VOLUME 156 OCTOBER 1988

TABLE

TABLE

I

II

Absolute Lymphocyte Count (cells/mm 3)

Time of Fistula Presentation Group

Patients

I

Postop Day

Group

Preop

At Fistulization

p Value

10,4 9 8.8 8.8 9.1 *

I & II III & IV p value

1,257 2,103 <0.01

1,024 1,467 <0.05

<0.05 <0.005

9

II Ill IV Overall

17 6 13 55

TABLE III

9 Average day.

Fistula Closure With Nonoperative Therapy

py and resection (Group IV). Initial fistula management in all cases was enteral feeding, antibiotics, and local wound care. Failure of this expectant therapy was followed by a variety of surgical procedures.

Group

I

TABLE

3/9

II III IV

RESULTS The onset of fistulization was similar in each group with an average time of presentation of 9.1 days after resection (Table I). Hemoglobin and albumin levels, absolute lymphocyte counts, and liver function test results were obtained both ,at the time of resection and at the time of fistulization. Liver function analysis revealed no abnormalities in each group both preoperatively and postoperatively. There was no significant difference in the hemoglobin and albumin levels among the groups. However, there was a significant decrease in the preoperative hemoglobin level from 13.7 to 11.1 g/dl (p <0.001) and in the albumin level from 4.1 to 3.1 g/dl (p <0.001) at the time of fistulization. The absolute lymphocyte counts were significantly lower in those receiving preoperative radiotherapy (Groups I and II) both at the time of surgery (p <0.01) and at the time of fistula presentation (p <0.05) when compared with those not receiving radiotherapy (Groups III and IV). The drop in lymphocyte count between surgery and fistula presentation was significant in all groups: Groups I and II, p <0.05; Groups III and IV, p <0.005 (Table II). Ninety-two percent of all fistulas in the study were eventually closed successfully. Utilizing nonoperative therapy, fistulas closed spontaneously in 3 of 9 patients in Group I, 10 of 17 patients in Group II, 13 of 16 patients in Group III, and 4 of 13 patients in Group IV. The average time for closure is illustrated in Table III. The remainder

Days After I:istulization

Patients

46 34 27 34

10/17 13/16 4/13

of the patients, except one in Group IV who died 51 days after fistula appearance, underwent attempted operative correction an average of 103, 26, 12, and 13 days after presentation (Table IV). Fistulas were repaired by debridement and primary approximation, advancement of existing musculocutaneous flap, or by reintroduction of fresh tissue. The methods of operative closure used were as follows: advancement of flap in 17 patients, pectoralis flap in 10 patients, primary flap in 4 patients, trapezius flap in 3 patients, sternocleidomastoid flap in 3 patients, jejunal flap in 2 patients, and latissimus flap in 2 patients. The final operative success rate was 83 percent. In Group I, attempted simple secondary closure was unsuccessful in one patient and the fistula persisted for 26 months. Four of the five patients treated with musculocutaneous flaps were ultimately cured of their fistulas in an average of 307 days (range 52 to 780 days) after development. One patient underwent multiple attempted closures (pectoralis, jejunum, and contralateral pectoralis), and the fistula remained open 22 months postoperatively. In Group II, 5 of 7 patients were approached with a second musculocutaneous flap. In 2 of 7 patients, a previously placed well-vascularized flap was advanced. Six of 7 patients had healed fistulas an average of 152 days postoperatively (range 1 to 630). One patient underwent

1V

Fistula Closure With Nonoperatlve Therapy*

Group

n

%

Attempted Closure (d)

Successful Closure (d)

Closure Method

Total Patients

Successful Closure

I

4/6

66

103

307

II

6/7

86

26

152

III

3/3

100

12

24

IV

7/8

88

13

27

Simple Flap Advanced flap Second flap Simple Flap Advanced flap Second flap

1 5 2 5 1 2 5 3

0 4 2 4 1 2 4 3

Patients

* Values for attempted closure and successful closure expressed as number of days after fistulizatien.

THE AMERICAN JOURNAL OFSURGERY

VOLUME 156

OCTOBER 1988

257

BROWN AND COLEMAN

a pectoralis flap with subsequent advancement and refistulization and was lost to follow-up three weeks after local advancement (210 days after original fistulization). In Group III, operative closure was successful in all three patients at an average of 24 days postoperatively (range 5 to 52 days) after fistula appearance. Methods used for closure were pectoralis major musculocutaneous flap (2 patients) and local reapproximation of tissue (1 patient). In Group IV, 5 of 8 patients were managed with local advancement of existing vascularized flap. Two were closed with a subsequent sternocleidomastoid musculocutaneous flap and one with a jejunal free flap. Seven of these patients were successfully closed an average of 27 days postoperatively after fistulization (range 6 to 72 days). One patient (local advancement) never healed and died 4 months after fistulization. Several serious complications were associated with the development of fistulas. Two patients in Group IV had mandibulotomy infections and nonunion after fistulas presented adjacent to the symphyseal closure. The fistulas healed 25 and 36 days after presentation without surgery, but the patients required treatment of osteomyelitis by debridement and external fixation, which was successful in both cases. Two patients experienced major arterial ruptures. One had been treated previously with laryngectomy and neck dissection for Stage IV carcinoma of the larynx with adjuvant radiotherapy (4,860 rads). Seven months later, mediastinal recurrence developed. It was resected and reconstruction was carried out with a jejunal free flap and pectoralis major muscle flap interposed between the trachea and innominate artery. Fistula at the distal anastomosis occurred on the ninth postoperative day and was followed by rupture of the innominate artery ! 0 days later. Ligation resulted in left hemiparesis and death within 4 days. The second patient had undergone prior radical neck dissection and adjuvant radiotherapy for an unknown primary tumor. He later underwent total glossectomy, laryngectomy, and free jejunal reconstruction. Mandibular recurrence 6 months later was treated by resection and pectoralis flap coverage. The fistula presented on the fifth postoperative day and was repaired 8 days later by debridement and advancement of local preexisting jejunum, healing without problems. Three months later, a second recurrence was resected and closed with a second pectoralis musculocutaneous flap. Fistula and rupture of the common carotid artery occurred on the third postoperative day. Ligation resulted in no neurologic deficit. Subsequent surgery for persistent fistula and sternal osteomyelitis included bilateral latissimus dorsi, trapezius, and rectus abdominis musculocutaneous flaps, which resulted in closure of the fistula. COMMENTS Preoperative radiotherapy, severe malnutrition, anemia, systemic disease, the presence of residual tumor, and the performance of radical neck dissection have all been implicated as etiologic factors in the development of fistulas. We examined the role of radiotherapy and musculo258

cutaneous flaps in the development and management of these fistulas and noted significant changes in several parameters of malnutrition. Chronic systemic disease and abnormal liver function test results were not prevalent in this population at the time of presentation. Preoperative hemoglobin and albumin levels were within normal limits for all groups. There was no difference in the groups with regard to radiotherapy. However, we found that the hemoglobin and albumin levels at the time of fistulization were significantly lower than preoperative values (p <0.001). This finding has also been reported by Lavelle and Maw [2], Wei et al [4], and Horgan and Dedo [5] who found a significant increased incidence of fistulas in patients with postoperative hemoglobin levels of less than 12.5, 12, and 11.5 g/ 100 ml, respectively. Although it is tempting to postulate that restoration of a normal hemoglobin level would aid in the prevention of fistulas, there is no conclusive evidence that normovolemic anemia in human subjects interferes with wound healing [6]. A decrease in serum albumin levels after surgery is not uncommon and can be explained by hemodilutional factors. In our study, preoperative serum albumin levels were within normal limits. All of our patients were placed on postoperative enteral feedings. The role of preoperative parenteral nutrition in this group is yet to be defined conclusively. Parenteral nutrition has been found by some investigators to decrease the incidence of postoperative complications in patients who were nutritionally depleted preoperatively but appeared to be of little benefit in patients with adequate preoperative nutrition [7-9]. Cell-mediated immunodeficiency is common in irradiated head and neck cancer patients and may contribute to fistula ocurrence [10-13]. Gray et al [11] found a significant difference in lymphocyte counts (p = 0.007) and corresponding T-cell, helper cell, and suppressor cell counts, and a decreased helper-to-suppressor cell ratio in patients with head and neck cancer who had undergone radiotherapy when compared with control subjects. They also demonstrated an additive decrease in lymphocyte counts in patients who underwent radiotherapy and surgery. In our study population, we found that previously irradiated patients (Groups I and II) had significantly lower absolute lymphocyte counts (p <0.01) prior to index surgery than nonirradiated patients (Groups III and IV). Moreover, there was a significant decrease in the lymphocyte count for each group from preoperative levels to levels at the time of fistulization (Groups I and II, p <0.05; Groups III and IV, p <0.005). This suggests that regional radiotherapy adversely affects the immune status of patients with head and neck carcinoma and the immune status is lowered in patients who develop fistulas. Some investigators have suggested an increased incidence of fistulization in patients who had undergone prior radiotherapy and radical neck dissection [2,14]. Sixtynine percent of our patients who developed fistulas had a concomitant radical neck dissection. Of these, only 32 percent had received prior radiotherapy. Perhaps the higher percentage of patients who underwent radical neck dissection and then developed fistulas is more a

THE AMERICANJOURNAL OFSURGERY VOLUME156 OCTOBER1988

RADIOTHERAPY~ MUSCULOCUTANEOUS FLAPS~ AND OROPHARYNGOCUTANEOUS FISTULAS

reflection of the higher stage disease at resection (82 percent with stage III and IV disease) than the presence of radical neck dissection with or without radiotherapy. Operations before the index resection did not seem to affect the eventual outcome of fistula healing. Twelve patients underwent such procedures. When these patients had recurrence, all but one underwent resection and closure with a musculocutaneous flap. Of these, seven conservatively treated patients had healing. Eleven percent (6 of 55 patients) had positive margins on final pathologic examination, half of which were considered to be a factor in the cause of fistulization. This compares with the 10 percent rate of residual tumor found as a cause of fistulization complicating total laryngectomy [4]. Several studies have associated preoperative radiotherapy with an increased incidence of orocutaneous fistulas [1,3,15,16], whereas others have found no such correlation [17,18]. From our data, we cannot comment on radiotherapy as a risk factor for fistulization but believe that it has a major influence on subsequent management. In those patients receiving radiotherapy who did develop fistulas, their course was longer and more severe. Major arterial ruptures are a risk in patients with oropharyngocutaneous fistulas who have received prior radiotherapy. Stell [19] has indicated an increased risk of carotid artery rupture in previously irradiated patients. In his series, 70 of 280 patients undergoing major neck surgery had been previously irradiated. Of these 70 patients, 9 suffered a major vessel rupture for an incidence of 13 percent in the irradiated patients. Orocutaneous fistulas are also implicated in the irradiated patient as increasing the risk of vessel rupture. In 40 of 63 carotid artery ruptures reported by Heller and Strong [20] and in 5 of 10 reported by Coleman [21], the presence of fistulas was considered a contributing factor. In our study, 2 of 22 patients who had received cancericidal radiotherapy experienced major rupture, for an incidence of 9 percent. We found that the musculocutaneous flap plays an invaluable role in the management of fistulas. Its presence in the irradiated patient at initial occurrence of a fistula is an important factor in eventual healing. In patients with no nonirradiated tissue (Group I), fistulas were more severe and prolonged (mean 307 days) than in those patients in whom flap closure was originally used (Group II, 152 days). Management in Group I was successful only by the introduction of newly vascularized extracervical tissue. Closure in Group II was accomplished by local advancement of an existing flap (29 percent) or by introduction of a second flap (71 percent). In the irradiated patient, the presence of a musculocutaneous flap was the only independently significant factor in the ability of the fistula to heal spontaneously. Differences in absolute lymphocyte count and hemoglobin and albumin levels between Groups I and II were not significant. A larger proportion of patients in Group II had closure without surgery and at a faster rate (Group II, 59 percent in 34 days and Group I, 33 percent in 46 days). The fistulas that presented in the nonirradiated patients were less severe and healed faster. In these patients,

nonsurgical and surgical management were more likely to be successful (Tables III and IV). There is a strong probability of spontaneous closure with conservative management up to 4 to 6 weeks after fistulization. If this fails to occur, patients should undergo operative closure. In Group IV, only a small fraction of fistulas actually closed spontaneously, undergoing operative correction instead after only 13 days. It is possible that many of these fistulas would have closed spontaneously if conservative management was pursued. Conversely, fistulas in Group I patients were permitted to close during a 3 month period before operative closure was undertaken. In all cases, spontaneous closure occurred within 6 weeks. Closure might have been expedited by attempting operative closure earlier within this time period. Many operative and nonoperative methods for the successful closure of oropharyngocutaneous fistulas have been described [22-25]. Ariyan [26] and others have advocated the use of pectoralis major muscle for reconstruction of head and neck defects. This and other vascularized pedicle flaps have been used extensively for reconstruction of head and neck defects and management of fistulas. In the present series, the pectoralis musculocutaneous flap was used in the majority of patients for initial mucosal closure and subsequent management of fistulas. Our operative success rate of 83 percent compares favorably with that of Rees et al [27], who successfully utilized musculocutaneous or muscle flaps to close fistulas in 88 percent of patients studied.

REFERENCES 1. Robbins RE, Budden MK, MacDougall JA. Analysis of mortality and morbidity in 100 composite resections for oral carcinoma. Am J Surg 1975; 130: 178-81. 2. Lavelle R J, Maw RA. The aetiology of post-laryngectomy pharyngocutaneous fistulae. J Laryngol Otol 1972; 86: 785-93. 3. Dedo DD, Alonso WA, Ogura JH. Incidence, predisposing factors and outcome of pharyngocutaneous fistulas complicating head and neck cancer surgery. Ann Otol Rhinol Laryngol 1975; 84: 833-40. 4. Wei WI, Lam KH, Wong J, Ong GB. Pharyngocutaneous fistula complicating total laryngectomy. Aust NZ J Surg 1980; 50: 366-9. 5. Horgan EC, Dedo HH. Prevention of major and minor fistulae after laryngectomy. Laryngoscope 1979; 89: 250-60. 6. Peacock EE. Wound repair. Philadelphia: WB Saunders, 1984:126. 7. Smale BF, Mullen JL, Buzby GD, Rosato EF. The efficacy of nutritional assessment and support in cancer surgery. Cancer 1981; 47: 2375-81. 8. Heatley RV, Williams RHD, Lewis MH. Preoperative intravenous feeding: a controlled trial. Postgrad Med J 1979; 55: 541-5. 9. Muller JM, Dienst C, Brenner U, Pichlmaier H. Preoperative parenteral feeding in patients with gastrointestinal cancer. Lancet 1982; 1: 68-71. 10. Olivari A, Pradier R, Feierstein J, Guardo A, Glait H, Rojas A. Cell-mediated immune response in head and neck cancer patients. J Surg Oncol 1976; 8: 287-94. 11. Gray WC, Chretien PB, Sutter CM, et al. Effects of radiation therapy on T-lymphocyte subpopulations in patients with head and neck cancer. Otolaryngol Head Neck Surg 1985; 93: 650-60. 12. Papenhausen PR, Kukwa A, Croft CB, Borowiecki B, Silver C, Emeson EE. Cellular immunity in patients with epidermoid cancer of the head and neck. Laryngoscope 1979; 89: 538-49. 13. Olkowski ZL, Wilkins SA. T-lymphocyte levels in peripheral blood of patients with cancer of the head and neck. Am J Surg 1975; 130: 440-4. 14. Briant TD. Spontaneous pharyngeal fistula and wound infection

THE AMERICAN JOURNAL OF SURGERY

VOLUME 156 OCTOBER 1988

259

BROWN AND COLEMAN

following laryngectomy. Laryngoscope 1975; 85: 829-34. 15. Mantavade RVP, Skolnik EM, Applebaum EL. Complications of postoperative and preoperative radiation therapy in head and neck cancers. Arch Otolaryngol Head Neck Surg 1981; 107: 690-3. 16. Joseph DL, Shumrick DL. Risks of head and neck surgery in previously irradiated patients. Arch Otolaryngol Head Neck Surg 1973; 97: 381-4. ! 7. Weingrad DN, Spiro RH. Complications after laryngectomy. Am J Surg 1983; 146: 517-20. 18. Bresson K, Rasmussen H, Rasmussen PA. Pharyngo-cutaneous fistulas in totally laryngectomized patients. J Laryngol Otol 1974; 88: 834-42. 19. Stell PM. Catastrophic hemorrhage after major neck surgery. Br J Surg 1969; 56: 525-7. 20. Heller KS, Strong EW. Carotid arterial hemorrhage after radical head and neck surgery. Am J Surg 1979; 138: 607-10.

260

THE AMERICAN JOURNAL OFSURGERY

21. Coleman JJ. Treatment of the ruptured or exposed carotid artery. South Med J 1985; 78: 262-7. 22. Myers EN. The management of pharyngocutaneous fistula. Arch Ototaryngol Head Neck Surg 1972; 95: 10-7. 23. Bellinger CG. Classification of pharyngostomes: a guideline for closure. Plast Reconstr Surg 1971; 47: 54-60. 24. Monihan RM, Lipshutz H. Method for closure ofa pharyngoeutaneous fistula. Plast Reconstr Surg 1971; 47: 384-5. 25. Vogel DH, Strong MS. Use of a T-tube in management of a pharyngeal fistula after laryngectomy. Plast Reconstr Surg 1978; 62: 573-5. 26. Ariyan S. The pectoralis major myocutaneous flap. Plast Reconstr Surg 1979; 63: 73-81. 27. Rees R, Cary A, Shack RB, Harris PF. Pharyngocutaneous fistulas in advanced cancer: closure with musculocutaneous or muscle flaps. Am J Surg 1987; 154: 381-3.

VOLUME 156 OCTOBER 1988