Abstract
Pharyngocutaneous fistula (PCF) is not uncommon after laryngeal surgery. Because it leads to prolonged hospitalization and increased patient morbidity, it is important to predict the formation of fistula. We studied the wound amylase concentrations of 87 consecutive laryngectomized patients. The PCF rate was 11.5% in this series. The amylase levels in patients with PCF begin to elevate significantly on the third, fourth, and fifth postoperative days. Our data indicate that increase in the wound amylase levels seems to be a significant predictor of PCF, and this should alert the surgeon for the presence of fistula.
Early detection of a fistula permits early intervention, and this will decrease morbidity and duration of hospital stay. Various methods are proposed to be effective in the early diagnosis of PCF 6-8 in the head and neck region. Amylase levels were studied for the prediction of pancreatic fistula in patients with total gastrectomy and PCF in patients with head and neck cancer, 9,10 but there have been no recent reports. In this study we aimed to find an early warning for the detection of PCF in the larynx and therefore, the amylase levels in the wound drainage were investigated.
MATERIALS AND METHODS
Eighty-seven patients requiring suction drainage after head and neck surgery are included in this study. The patients compose a consecutive series of laryngectomies, conducted at our clinic. All patients had prophylactic antibiotics and all had identical suture material (Vicryl) used for closure. Running Connell-type suture was used in a single vertical line to close the pharynx. A second layer of horizontal interrupted sutures approximated the 2 inferior constrictor muscles and their upper edge to the base of tongue.
Suction drains were placed at the time of the operation. All patients had serial estimation of their wound amylase concentration until the suction and drainage tubes were taken out. The wound drainage fluid was collected in sterile suction bottles and changed twice daily for samples to be sent to the laboratory. P-nitrophenol enzymatic calorimetric technique with autoanalyzer (Olympus 5200, Japan) was used to measure the amylase levels. The normal limits for amylase were 0 to 90 IU/L in our laboratory. If there is an increase in the concentration of wound amylase, oral feeding was delayed; otherwise it was started on the third to fifth postoperative days depending on the judgment of the surgeon.
We also studied other factors such as age, preoperative radiotherapy, and T stage of the lesions.
For statistical analysis, SPSS 9.0 Windows Software Package was used. Two-sided general linear univariate model Tukey post hoc and Dunnett tests and t tests were used for intergroup comparison. Pearson's correlation test was used to assess the correlation of PCF formation, and t tests were used for equality of means.
Comparison of characteristics of patients with and without PCF
RESULTS
Of 87 patients, 84 (96.5%) were men and 3 (3.5%) were women. The average age was 61 years, with a range of 38 to 79 years. There were 33 patients with T3 lesions, 49 patients with T4 lesions, and 5 patients with T2 lesions who were primary radiation failures. The final surgical pathology was squamous cell carcinoma in all of the patients.
The postoperative PCF rate was 11.5%. The mean age of patients with and without PCF was 61.9 (SD ± 12.1), and 56.9 (SD ± 9.8), respectively. When the mean ages of these 2 groups were compared, there was no statistically significant difference (P > 0.05). Six patients (7%) had preoperative radiation therapy. Of these patients, only 2 (33.3%) developed PCF. On the other hand, 8 patients (10.3%) who did not receive radiation therapy preoperatively went on to have PCF. The statistical analysis did not reveal any significant difference between these 2 groups for the PCF formation (P > 0.05).
In addition, PCF developed in 2 of 33 patients with T3 lesions and 8 of 49 patients with T4 lesions. These data were not statistically significant (P > 0.05). The results of these variables are summarized in Table 1.
The mean wound amylase concentrations of all patients on postoperative days 1 through 5 are given in Table 2. Analysis of these data showed that the wound amylase levels did not increase significantly in the postoperative period (P > 0.05). The mean amylase levels were grouped according to the presence of PCF for each postoperative day. The mean values of these 2 groups were compared for each day. There was no statistically significant difference between the amylase levels of these 2 groups of patients on the postoperative days 1 and 2 (P > 0.05), but the differences became statistically significant on the postoperative days 3 through 5 (P < 0.05) (Table 3).
Mean wound amylase levels in relation to postoperative days
When the wound amylase levels of patients with fistula on postoperative days 1 through 5 were compared, the statistically significant differences were noted between the postoperative days 1 versus 4, 1 versus 5, 2 versus 4, and 2 versus 5 (P < 0.05). These data also support that wound amylase concentrations elevate significantly. On the other hand, wound amylase levels of patients without PCF decreased dramatically beginning from postoperative day 3.
There were no statistically significant differences when all parameters were compared with each other (P < 0.05).
DISCUSSION
Most fistulas occur early in the postoperative period and close spontaneously, but some require multiple surgical procedures. PCF leads to a complicated rehabilitation and lengthens the patient's stay in the hospital more commonly than any other postoperative complications. 5 Various studies have been conducted to predict fistula as early as possible. Some of these studies have managed to diagnose PCF within an average of 5 to 10 days. 7,11 To predict the salivary fistula, Krouse and Metson 6 performed barium swallows in 51 of 109 consecutive laryngectomized patients 1 to 2 weeks after the operation. The presence of a sinus tract 2 cm or longer on barium swallow was regarded as a predictor of a postlaryngectomy PCF. The authors concluded that barium swallow was an excellent predictor of the development of a clinical salivary fistula.
In another study, the role of cinepharyngoesophagogram was examined to evaluate the neopharynx pharynx in the early postoperative period in total laryngectomy patients. 7 The authors of this retrospective study suggested that when there were no clinical signs or symptoms suggestive of an impending fistula, there was no need to perform a cinepharyngoesophagogram before starting oral alimentation.
Comparison of postoperative amylase levels in patients with and without PCF
Friedman et al 8 studied the fever (>101.5° F) in the first 48 hours after the surgery as an indicator of PCF and suggested that fever was an excellent tool for early identification of PCF.
In this study we examined the relevance of wound amylase concentration in a series of laryngeal cancer patients for early detection of PCF. Amylase is an enzyme present in small amounts in serum but in large quantities in saliva. Amylase contamination of wound drainage after head and neck surgery occurs when either salivary tissue has been transected or the oral cavity has been opened. When neither has occurred, wound amylase is derived entirely from the serum and levels are low.
The absolute concentration of amylase in wound drainage after head and neck surgery can be measured easily. Larsen and Schuller 9 showed that when wound healing is uneventful, the amylase concentration in the drainage falls gradually postoperatively. They were able to predict very early fistulas by observing a rising wound amylase level.
Fielder and Morton 5 suggested that it was possible to predict fistula formation as early as the second postoperative day, using serial measurements of the amylase concentration in wound drainage.
In our study, a gradual decrease in the amount of wound amylase levels was observed postoperatively. Amylase levels of patients with and without PCF on the postoperative days 1 and 2 did not reveal any significant difference, but there were statistically significant differences between the amylase levels on the postoperative days 3, 4, and 5 (P < 0.05). The results of this study were in accordance with the previous studies. 5,9 These data suggest that an increase in the wound amylase concentrations can be a useful marker in the early identification of PCF, thus leading to early rehabilitation. As a cheap and reliable marker, wound amylase level is easy to obtain as long as the drains left in place. We prefer to keep the drains in place until they are dry. Because a majority of the drains were kept in place during the first 5 postoperative days, serial measurements were performed until they were removed.
In conclusion, serial estimation of wound amylase levels is a simple and cheap test for predicting the development of a clinical salivary fistula.