Abstract
Purpose
The purpose of this study was to determine the incidence of and risk factors for pharyngocutaneous fistula in patients undergoing total laryngectomy at a single institution.
Materials and methods
The records of 59 patients undergoing primary or salvage total laryngectomy at our institution from 2001 to 2012 were retrospectively reviewed. Data collected included patient, tumor and treatment characteristics, and surgical technique. Risk factors were analyzed for association with pharyngocutaneous fistula formation.
Results
Twenty patients (34%) developed fistulas. Preoperative tracheostomy (OR 4.1; 95% CI 1.3–13 [ p = 0.02]) and low postoperative hemoglobin (OR 9.1; 95% CI 1.1–78 [ p = 0.04]) were associated with fistula development. Regarding surgical technique, primary sutured closure of the total laryngectomy defect had the lowest fistula rate (11%). In comparison, primary stapled closure and pectoralis onlay flap over primary closure had nonsignificantly increased fistula rates (43%, OR 6.0; 95% CI 1.0–37.3 [ p = 0.06] and 25%, OR 2.7; 95% CI 0.4–23.9 [ p = 0.38], respectively). Pectoralis flap incorporated into the suture line had a significantly increased fistula rate (50%, OR 7.1; 95% CI 1.4–46 [ p = 0.02]). After stratification for salvage status, patient comorbidities were associated with fistula in non-salvage cases whereas disease-related characteristics were associated with fistula in salvage cases. Fistula development was associated with increased length of hospital stay ( p < 0.001) and increased time before oral diet initiation ( p < 0.001).
Conclusions
Pharyngocutaneous fistula is a common complication of total laryngectomy. Preoperative tracheostomy, postoperative hemoglobin, and surgical technique are important in determining the risk of fistula.
1
Introduction
Total laryngectomy (TL) carries with it significant morbidity, with 40%–92% reported complication rates . Complications include wound infection and dehiscence, swallowing dysfunction, airway complications, chyle leak, and carotid artery rupture . The most frequent postoperative complication is the development of a pharyngocutaneous fistula (PCF), which occurs with variable frequencies of 3% to 65% in reported series, most falling within 10%–40% . PCF is an early complication that causes increased length of hospital stay and delayed initiation of oral diet, and requires complex wound care, occasionally necessitating additional surgery for closure .
Case series have identified many factors as significantly associated with PCF development, although consistency across studies is lacking. Preoperative radiation therapy increases both the frequency and severity of postoperative PCF , and chemotherapy augments this effect . Other factors associated with increased rates of PCF include patient comorbidities , hypothyroidism , low perioperative hemoglobin and albumin , more advanced primary tumor stage , prior tracheotomy positive surgical margins , concurrent neck dissection , shorter elapsed time from completion of radiation to surgery , and surgical closure technique . Although it is generally accepted that previous radiation and chemoradiation increase the risk of PCF , there is little consensus regarding the relative significance of these other various factors. In the era of organ preservation therapy with TL often reserved as a salvage measure after failed radiation with or without chemotherapy, recent efforts have largely focused on identifying surgical techniques to minimize PCF formation in tissue that is frequently irradiated and at high risk for poor wound healing .
The aims of this study were to review our single-center experience with PCF after TL and to determine risk factors associated with PCF in our patient population.
2
Material and methods
2.1
Patients
Approval for this study was obtained from the Greater Baltimore Medical Center (GBMC) Institutional Review Board. The records of 59 patients who underwent TL at GBMC in Baltimore, MD from 2001 to 2012 were retrospectively reviewed. Of note, 19 of these patients (study ID numbers 3, 4, 11, 14–18, 20, 21, 24, 28–34, 36) were also included in a previously reported multicenter study of PCF after salvage total laryngectomy . Patient, disease, and treatment data were collected. Patient data included age, gender, comorbidities, social history (tobacco and alcohol use) and laboratory data. Comorbidities were scored using the Washington University Head and Neck Cancer Comorbidity Index (WUHNC Index) .
2.2
Disease
The indication for surgery (primary versus salvage treatment, and cancer versus dysfunctional larynx), tumor site, and disease stage at the time of treatment were recorded.
2.3
Treatment
Pre- and postoperative radiation and chemotherapy were recorded. Salvage was defined as having received prior radiation therapy with or without chemotherapy. Surgical treatment data included any previous surgical treatment, surgical technique, any extent of pharyngectomy, concurrent neck dissection, preoperative tracheostomy, and margin status.
2.4
Outcomes
The primary outcome was PCF, defined as a frankly draining salivary tract from the neopharynx to the skin requiring either wound packing or surgical management. Additional outcomes included length of hospital stay, days before initiation of oral diet, and survival.
2.5
Statistical analysis
Descriptive statistics were reported as n (%) for categorical variables and as mean and standard deviation (SD) or median and range for continuous variables. Categorical variables as predictors of PCF occurrence were analyzed using the Fisher’s exact test and logistic regression. Laboratory data were dichotomized based on our center’s laboratory criteria and analyzed with logistic regression, and/or considered a continuous variable and analyzed using the Mann–Whitney U test. Differences in median length of stay and time before oral diet by PCF status were analyzed using a nonparametric equality-of-medians test. Multivariable analyses were performed using logistic regression with varying subsets of predictors based on the univariate analyses and/or previous literature. Analysis of PCF formation was performed for the entire cohort, after stratification for salvage, and with exclusion of pharyngectomy cases. Survival analysis was performed using the Kaplan–Meier method. Data analysis was performed using commercially-available software (STATA, version 11.2; StataCorp LP, College Station, TX, 2012). A p-value ≤ 0.05 was considered statistically significant.
2
Material and methods
2.1
Patients
Approval for this study was obtained from the Greater Baltimore Medical Center (GBMC) Institutional Review Board. The records of 59 patients who underwent TL at GBMC in Baltimore, MD from 2001 to 2012 were retrospectively reviewed. Of note, 19 of these patients (study ID numbers 3, 4, 11, 14–18, 20, 21, 24, 28–34, 36) were also included in a previously reported multicenter study of PCF after salvage total laryngectomy . Patient, disease, and treatment data were collected. Patient data included age, gender, comorbidities, social history (tobacco and alcohol use) and laboratory data. Comorbidities were scored using the Washington University Head and Neck Cancer Comorbidity Index (WUHNC Index) .
2.2
Disease
The indication for surgery (primary versus salvage treatment, and cancer versus dysfunctional larynx), tumor site, and disease stage at the time of treatment were recorded.
2.3
Treatment
Pre- and postoperative radiation and chemotherapy were recorded. Salvage was defined as having received prior radiation therapy with or without chemotherapy. Surgical treatment data included any previous surgical treatment, surgical technique, any extent of pharyngectomy, concurrent neck dissection, preoperative tracheostomy, and margin status.
2.4
Outcomes
The primary outcome was PCF, defined as a frankly draining salivary tract from the neopharynx to the skin requiring either wound packing or surgical management. Additional outcomes included length of hospital stay, days before initiation of oral diet, and survival.
2.5
Statistical analysis
Descriptive statistics were reported as n (%) for categorical variables and as mean and standard deviation (SD) or median and range for continuous variables. Categorical variables as predictors of PCF occurrence were analyzed using the Fisher’s exact test and logistic regression. Laboratory data were dichotomized based on our center’s laboratory criteria and analyzed with logistic regression, and/or considered a continuous variable and analyzed using the Mann–Whitney U test. Differences in median length of stay and time before oral diet by PCF status were analyzed using a nonparametric equality-of-medians test. Multivariable analyses were performed using logistic regression with varying subsets of predictors based on the univariate analyses and/or previous literature. Analysis of PCF formation was performed for the entire cohort, after stratification for salvage, and with exclusion of pharyngectomy cases. Survival analysis was performed using the Kaplan–Meier method. Data analysis was performed using commercially-available software (STATA, version 11.2; StataCorp LP, College Station, TX, 2012). A p-value ≤ 0.05 was considered statistically significant.
3
Results
3.1
Patient characteristics
This study population included 59 patients with a mean age of 63 years (standard deviation, 10.6; range 38–89 years). The median follow-up time for all patients was 19.6 months (range, 1.2–124.1 months). For the 54 patients whose primary indication for TL was malignancy, median overall survival was 33.3 months (standard error [SE] 3.7) and median disease-specific survival was 42.2 months (SE 9.0). Thirty-seven patients (63%) were considered salvage cases, and median time from completion of radiation to TL was 15 months (range, 2 to 112 months).
TL was performed using various surgical techniques including: Primary closure with suture (18 patients, 31%); primary closure with stapler (14 patients, 24%); PM myofascial onlay flap (PM onlay) over either sutured or stapled primary closure (eight patients or 14%, including one sutured closure and seven stapled closures); PM myocutaneous flap incorporated into the suture line (PM in suture line) (18 patients, 31%); and radial forearm free tissue transfer (one patient, 2%).
Patient characteristics were compared among the different closure groups. Statistically different characteristics are shown in Table 1 . Compared with other techniques, a greater proportion of PM in suture line closures were used in cases with pharyngectomy defects (74% PM in suture line versus 0–17% of others, p < 0.001), and a greater proportion of PM in suture line and PM onlay closures were employed in salvage cases (89% PM in suture line and 88% PM onlay versus 43%–44% of others, p = 0.004).
| Closure b | N | Primary with suture, N (%) | Primary with stapler, N (%) | PM onlay, N (%) | PM in suture line, N (%) | p -Value c |
|---|---|---|---|---|---|---|
| Total N | 59 | 18 | 14 | 8 | 19 | |
| Characteristic | ||||||
| HTN | 0.025 | |||||
| No | 28 | 6 (33) | 4 (29) | 7 (88) | 11 (58) | |
| Yes | 31 | 12 (67) | 10 (71) | 1 (12) | 8 (42) | |
| Pulmonary disease | 0.025 | |||||
| No | 42 | 14 (78) | 6 (43) | 5 (63) | 17 (89) | |
| Yes | 17 | 4 (22) | 8 (57) | 3 (38) | 2 (11) | |
| Obesity | 0.024 | |||||
| No | 52 | 15 (83) | 13 (93) | 5 (63) | 19 (100) | |
| Yes | 7 | 3 (17) | 1 (7) | 3 (38) | 0 (0) | |
| Disease | ||||||
| Salvage | 0.004 | |||||
| No | 21 | 10 (56) | 8 (57) | 1 (13) | 2 (11) | |
| Yes | 38 | 8 (44) | 6 (43) | 7 (88) | 17 (89) | |
| Site | < 0.001 | |||||
| Glottic | 28 | 9 (50) | 12 (86) | 4 (50) | 3 (16) | |
| Supraglottic | 14 | 8 (44) | 1 (7) | 3 (38) | 2 (11) | |
| Hypopharyngeal | 10 | 0 (0) | 0 (0) | 0 (0) | 10 (53) | |
| Oropharyngeal | 5 | 0 (0) | 0 (0) | 1 (12) | 4 (21) | |
| Other d | 2 | 1 (6) | 1 (7) | 0 (0) | 0 (0) | |
| Pharyngectomy | < 0.001 | |||||
| No | 42 | 15 (83) | 14 (100) | 8 (100) | 5 (26) | |
| Yes | 17 | 3 (17) | 0 (0) | 0 (0) | 14 (74) | |
| Positive margins | 0.037 | |||||
| No | 39 | 13 (72) | 13 (93) | 4 (50) | 9 (50) | |
| Yes | 20 | 5 (28) | 1 (7) | 4 (50) | 9 (50) |
a This table includes variables that showed significant differences across the closure categories. Other variables examined that did not show significant associations include: age, gender, tobacco use, alcohol use, diabetes, coronary artery disease, Washington University Head and Neck Comorbidity Index, preoperative hypothyroidism, preoperative hypoalbuminemia, preoperative hemoglobin < 12.5, postoperative hemoglobin < 12.5, indication for total laryngectomy (cancer versus dysfunctional larynx), preoperative tracheostomy, and neck dissection.
b The single patient undergoing free flap closure was excluded from these analyses.
c Significance calculated using Fisher’s exact t-test.
d “Other” includes patients with unknown primary and no malignancy (dysfunctional larynx).
3.2
Pharyngocutaneous fistula
Twenty patients (34%) developed PCF. With univariate analysis ( Table 2 ), postoperative hemoglobin less than 12.5 g/dL was significantly associated with PCF (OR 9.1; 95% CI 1.1–78 [ p = 0.04]). Preoperative hypothyroidism with TSH > 4.7 mIU/L perfectly predicted PCF formation, with 4 out of 4 patients developing PCF. Preoperative TSH was higher in patients who developed PCF compared to those who did not (median 5.0 mIU/L versus 0.7 mIU/L, p = 0.0023).
| Characteristic | N | No. patients (% of total patients) | No. patients with PCF (% of patients in category) | OR; 95% CI ( p ) a |
|---|---|---|---|---|
| Patient | ||||
| Age: (range 38–89, mean 63) | 59 | |||
| < 6 | 18 (31) | 6 (33) | REF | |
| ≥ 60 | 41 (69) | 14 (34) | 0.9; 0.3–3.4 (0.95) | |
| Gender: | 59 | |||
| Female | 11 (19) | 3 (27) | REF | |
| Male | 48 (81) | 17 (35) | 1.5; 0.3–6.3 (0.61) | |
| Social history: | 59 | |||
| Tobacco use | ||||
| No | 10 (17) | 1 (10) | REF | |
| Yes | 49 (83) | 19 (39) | 5.7; 0.7–50 (0.12) | |
| Alcohol use | ||||
| No | 31 (53) | 12 (39) | REF | |
| Yes | 28 (47) | 8 (29) | 0.63; 0.2–1.9 (0.41) | |
| Medical history: | 59 | |||
| DM | ||||
| No | 45 (76) | 14 (31) | REF | |
| Yes | 14 (24) | 6 (43) | 1.7; 0.5–5.8 (0.42) | |
| HTN | ||||
| No | 27 (46) | 6 (22) | REF | |
| Yes | 32 (54) | 14 (44) | 2.7; 0.9–8.6 (0.09) | |
| Pulmonary disease b | ||||
| No | 41 (69) | 12 (29) | REF | |
| Yes | 18 (31) | 8 (44) | 1.9; 0.6–6.1 (0.26) | |
| CAD | ||||
| No | 45 (76) | 13 (29) | REF | |
| Yes | 14 (24) | 7 (50) | 2.5; 0.7–8.5 (0.16) | |
| Obesity | ||||
| No | 51 (86) | 17 (33) | REF | |
| Yes | 8 (14) | 3 (38) | 1.2; 0.3–5.7 (0.81) | |
| WUHNC index | 59 | |||
| 0 | 34 (58) | 9 (26) | REF | |
| 1 + | 25 (42) | 11 (44) | 2.2; 0.7–6.6 (0.17) | |
| Laboratory data | ||||
| Preop hypothyroidism (TSH > 4.7 mIU/L) | 16 | N/A c | ||
| No | 12 (75) | 4 (33) | ||
| Yes | 4 (25) | 4 (100) | ||
| Preop albumin ≤ 3.2 g/dL | 42 | |||
| No | 39 (93) | 15 (38) | REF | |
| Yes | 3 (7) | 2 (67) | 3.2; 0.3–40 (0.37) | |
| Preop Hgb < 12.5 g/dL | 57 | |||
| No | 34 (60) | 10 (29) | REF | |
| Yes | 23 (40) | 10 (43) | 1.8; 0.6–5.6 (0.28) | |
| Postop Hgb < 12.5 g/dL | 57 | |||
| No | 13 (23) | 1 (8) | REF | |
| Yes | 44 (77) | 19 (43) | 9.1; 1.1–78 (0.04) | |
| Disease | ||||
| Indication for surgery | 59 | |||
| Cancer | 54 (92) | 17 (32) | REF | |
| Dysfunctional larynx | 5 (8) | 3 (60) | 3.3; 0.5–21.7 (0.22) | |
| Site | 57 | |||
| Glottic | 27 (46) | 9 (33) | REF | |
| Hypopharyngeal | 11 (19) | 5 (45) | 1.7; 0.4–7.1 (0.49) | |
| Supraglottic | 14 (24) | 3 (21) | 0.5; 0.1–2.5 (0.43) | |
| Oropharyngeal | 5 (8) | 2 (40) | 1.3; 0.2–9.6 (0.78) | |
| Staging | ||||
| T stage | 54 | |||
| T2 | 13 (24) | 2 (15) | REF | |
| T3 | 14 (26) | 4 (29) | 2.2; 0.3–15 (0.42) | |
| T4 | 27 (50) | 11 (41) | 3.8; 0.7–21 (0.13) | |
| Overall stage | 54 | |||
| II–III | 25 (46) | 5 (20) | REF | |
| III–IV | 29 (54) | 12 (41) | 2.8; 0.8–9.7 (0.10) | |
| Treatment | ||||
| Salvage | 59 | |||
| No | 22 (37) | 6 (27) | REF | |
| Yes | 37 (63) | 14 (38) | 1.6; 0.5–5.2 (0.41) | |
| Previous surgical treatment | 58 | |||
| No | 44 (76) | 13 (30) | REF | |
| Yes | 14 (24) | 7 (50) | 2.4; 0.7–8.3 (0.17) | |
| Closure | 59 | |||
| Primary with suture | 18 (31) | 2 (11) | REF | |
| Primary with stapler | 14 (24) | 6 (43) | 6.0; 1.0–37.3 (0.06) | |
| PM onlay | 8 (14) | 2 (25) | 2.7; 0.4–23.9 (0.38) | |
| PM in suture line | 18 (31) | 9 (50) | 7.1; 1.4–46 (0.02) | |
| Free flap d | 1 (5) | 1 (100) | ||
| Pharyngectomy e | 59 | |||
| No | 42 (71) | 14 (33) | REF | |
| Yes | 17 (29) | 6 (35) | 1.1; 0.3–3.6 (0.89) | |
| Neck dissection f | 59 | |||
| No | 20 (34) | 8 (40) | REF | |
| Yes | 39 (66) | 12 (31) | 0.67; 0.2–2.1 (0.48) | |
| Preop tracheostomy | 59 | |||
| No | 39 (66) | 9 (23) | REF | |
| Yes | 20 (34) | 11 (55) | 4.1; 1.3–13 (0.02) | |
| Positive margins | 59 | |||
| No | 39 (66) | 13 (33) | REF | |
| Yes | 20 (34) | 7 (35) | 1.1; 0.3–3.4 (0.90) | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
